CN101350023B - Method and platform capable of customizing enquiry based on service combination - Google Patents

Abstract

The invention provides a customizable query method and a platform on the basis of service composition. The method comprises: describing own demands to the platform by a user through an appointed client end visible component and through operations of selecting and filling the attribute value of the component, analyzing the initial state and the target state of a data query request, then, searching and combining a corresponding basic service calling sequence, then, executing basic services in the sequence, thereby obtaining a result set which contains user expected result data, and generating a final result to display to the user through screening, sequencing and assembling. In the method, the user describes own demands to a system actively, the system can effectively adapt different individualized demands of different users through receiving user description and dynamic analysis processing, the application system development work load and the later maintenance cost are lowered, and the friendliness of a data query application system is improved, which enables the user to customize data query actively and enables inquiry to be flexible and convenient.

Description

A Customizable Query Method and Platform Based on Service Composition

technical field

The invention belongs to the field of computer data management, and in particular relates to a service combination-based customizable data query method and a platform thereof.

Background technique

In the traditional web-based data query, the web application system provides an independent data query condition input page for each data query function. The query condition input page is the entrance of this data query function, and the background logic corresponding to the page is fixed and implemented in a system hard-coded way. For the function of inputting complex query conditions, the system provides an input method for simple query conditions such as "and", "or" and "not" to form complex query conditions.

When users query data, they first select the required function page from the function set provided by the system, then select and fill in the query conditions and submit the data query request according to the method specified by the system, and finally obtain the desired results in the fixed format response returned by the system. The system regards no difference between different users entering from the same data query condition input page, and provides the same query service for such users.

Such traditional methods have the following problems:

First, the fixed data query processing method cannot meet the user's personalized query needs. For the same function, the application system can only accept the input conditions provided by the page, and the system output is the data query result in a fixed format. If the query condition expected by the user is not the condition input page item provided by the system, the query requirement cannot be accepted by the system; if the user does not care about some data in the query result obtained by the system and wants it to be filtered for data analysis, Or if you want to focus on a certain part of the data, then the query processing method of the system with a fixed output result format cannot meet the needs of this user preference.

Second, the data query functions supported by the application system have been classified in advance, and a user's query action is limited to a page provided with the query function. If the query classification required by the user is different from the query function classification of the application system, the query requirement cannot be fulfilled on one query page provided by the system, and the user needs to perform multiple queries on other related system pages, and the user will logically Multiple query results obtained from the above comparison. This method leads to low data query efficiency, increased user workload, and poor system friendliness.

Third, when developing an application system, it is necessary to do hard-coding work for each data query function, which is a heavy workload and a lot of repetitive labor.

Fourth, after the data query application system is developed and put into use, if there are new functional requirements, the system cannot automatically adapt to the changes in requirements. If it is solved by system upgrade, it will consume manpower and material resources, and at the same time, it will easily destroy the original structure of the system, making the system gradually lose maintainability and scalability.

Patent No. 200610059824.1, "A Custom Query Method and System for Performance Data" proposes a custom query method for performance data, specifying one or more related attributes of the query object as template attributes through the system customization template and saving the template, the user Construct a query request by selecting a template and setting query parameters. The server parses out the template attributes and query parameters to be queried, and constructs a query statement to execute the database query. In this method, the user can only select the required query mode in the template provided by the system. The template specifies what kind of attribute data is the result of a query. The template customization is completed by the system administrator, and the user cannot directly intervene. When the system provides When the template cannot meet the user's query requirements, the user still needs to perform multiple queries and manually assemble the query results, and even cannot obtain the expected data through the system. The server-side query engine processes the user data query request by dynamically constructing an SQL statement according to the query parameters and then executing the database query. Before the SQL statement generated by this processing method is executed, its logical rationality cannot be verified, and a reasonable SQL statement cannot be associated with the corresponding query function, so it is not reusable.

Contents of the invention

The present invention proposes a customizable query method and platform based on service combination. Users can describe their needs, initial status and goals of data query requests to the platform through specified client-side visual components and operations of selecting and filling component attribute values. After the state is parsed, search and combine the corresponding basic service call sequence, and then execute the basic service in the sequence, and then get the result set containing the user's expected result data, and generate the final result after filtering, sorting, and assembling to display to the user . In this method, users actively describe their needs to the system, and the system can effectively adapt to different individual needs of different users by receiving user descriptions and dynamically analyzing and processing them, reducing the workload of application system development and later maintenance costs, and improving the efficiency of data query application systems. friendliness. It enables users to actively customize data query, making the query flexible and convenient.

A customizable query method based on service composition, comprising the following steps:

Step 1: Establish a customizable query client and a data query server, abstract and classify the data in the domain, build an editable control for the established model, and store it in the domain data object model library in the customizable query client. Build the corresponding basic data query service in the data query server and store it in the basic data query service library;

Step 2: Initialize the domain data object model library of the customizable query client and the template library that stores existing custom functions; the user compares his individual needs with the existing functions in the template library, and if there is a template library that meets the user's current needs Function template, select a specific function template from the template library, and the page display corresponds to the customized query content, query conditions, and result display format for this function, and adjust the display format, refill the customized query condition attribute value, and construct the same function The new request; otherwise, the domain data object model library and custom result encapsulation sub-module customize data query;

Step 3: You can customize the query client's customized results and encapsulate the submodule to encapsulate the user's customized results; if this customization comes from the template library, set the template's identification bit to indicate that it comes from an existing customized template, otherwise set the corresponding identification to indicate that it is new Customization; the request queue management submodule sent to the data query server;

Step 4: The request queue management submodule receives the data query request sent by the client and adds it to the request queue;

Step 5: The request parsing sub-module takes out the first request in the request queue, and obtains the initial state, target state, constraint conditions and result display format after parsing the request; if the request queue is not empty, judge whether the data query is a template library One of the functions, if it is, the execution sub-module directly obtains its service call sequence, and then go to step 8; otherwise, call the service composition engine to perform service composition; if the request queue is empty, go to step 10;

Step 6: According to the initial state and the target state, the service composition engine calls the basic data that can be used currently obtained by the service agent sub-module to query service information, and processes the query process in the initial state one by one. On the premise of satisfying the constraints of the intermediate process, for Each query process performs service combination to generate a service call sequence; if the service combination fails, go to step 10;

Step 7: Store the corresponding function and service call sequence into the template library by the customized storage module;

Step 8: The execution sub-module calls the basic data query service in the basic data query service library according to the order of the service call process sequence, and coordinates the input and output relationship between services; if there is an exception, go to step 10;

Step 9: The result generation sub-module generates the final data query and analysis result according to the result display format described in the request, returns a response to the client, and displays the result;

Step 10: The exception handling sub-module generates failure information and returns it to the client for display.

The abstract classification of the data in the step 1 is to extract the features of the data according to the object-oriented thinking; the attributes contained in the abstracted domain data object model are used as the attributes of the object model, and each domain data object model is based on the customer The client-side script is written as an editable control that can customize the query client, and describes its own data query requirements to the application system.

The domain data object model library in the step 1 includes logic controls, which describe the relationship between two or more domain data object attributes by means of operations and logic operations, and each logic control is written in a script.

The basic data query service of the basic data query service library in the step 1 corresponds to the domain data object model of the domain data object model library one by one, and provides the atomic operation of the domain data object query, only through one of the corresponding domain data object models or several attributes, and obtain a collection of data object instances with this attribute from the background database.

Each basic data query service of the basic data query service library in step 1 includes an atomic operation for querying associated attributes, an atomic operation for obtaining associated data objects and their attributes of corresponding domain data objects, and outputting associated data objects and associated data A collection of properties for an object.

The customized data query in the step 2 includes the following steps:

Step a: On the page of the customizable query client, select each editable control from the controls in the domain data model library, specify each query content data object of this query; select the properties of each editable control as the result of this data query A column of the data list;

Step b: Select each editable control from the visual controls of the domain data model library on the customization page, specify the conditional data object for this query; select the attribute of each editable control, select and fill in the attribute value for this attribute, and Specify the query content data object used for the query condition selected this time;

Step c: Arrange the sequence of the selected control properties, fill in the data list header of the custom page result, and specify the sorting column of the result list;

Step d: Name the newly customized data query, save this customization, and name the template corresponding to this customization.

The initial state in step 5 is composed of query process set, query result and service call sequence; the query process set includes the query process of the data object to be queried in this user request, and the query result and service call sequence are empty.

The target state in the step 5 is composed of query process set, query result and service call sequence; the query process set is empty, the query result is the target data object of each query process in the query process set, and the service call sequence is an idle period memory space.

The service combination in the step 6 includes the following steps:

Step1: The service composition engine extracts the source state of this service composition process, including the target data object of the query process, query condition set and empty service call sequence; add the source state to the available state queue for temporary storage of data;

Step2: If the available status queue is empty, then output service combination failure information, and the service combination process ends; otherwise, go to step3;

Step3: Take the available state at the head of the queue from the available state queue as the current state;

Step4: If the query condition of the current state is empty, output the service call sequence of the current state, and the service composition process ends; otherwise, store the query condition of the current state in the temporary queue for storing the query condition;

Step5: If the temporary queue is empty, go to step2; otherwise, go to step6;

Step6: Take the query condition at the head of the queue from the temporary queue as the current condition, search for the basic data query service that can be used for the current condition and find its corresponding atomic operation, if all the basic data query services have been called, go to step2 ;Otherwise go to step7;

Step7: Add the found atomic operation to the service call sequence of the current state in an equivalence relationship;

Step8: According to the current query condition of the current state, generate the successor state of the current state, and add the successor state to the available state queue, then go to step5.

When the step8 generates the successor state of the current state, if the domain data object corresponding to the current condition is the target object of the query process, delete the current condition from the current state, and add the state after deleting the current condition as the successor state to the available state Queue, generated; otherwise, call the associated query atomic operation of the basic data query service corresponding to the current condition, obtain the set of associated attributes, add each associated attribute and the source of the attribute value, and delete the current condition from the current state as a successor state Join the available state queue; then process the next associated attribute and add all subsequent states to the available state queue until all associated attributes in the associated attribute set have generated subsequent states.

A customizable query platform based on service composition, characterized in that it includes the following modules:

The established customizable query client and data query server; the customizable query client includes the domain data object model library, which is used to abstract and classify the data in the domain and store them in the editable control built for the established model; data query The server, including the basic data query service library, builds and stores the basic data query service corresponding to the editable control;

The customizable query client also includes a template library and a custom result encapsulation sub-module;

The template library is used to store the existing customized functions. After the template library is initialized, the user compares his individual needs with the existing functions in the template library. Select a template for a specific function, and the page display corresponds to the customized query content, query conditions, and result display format for this function, and adjust the display format, refill the custom query condition attribute value, and construct a new request for the same function; otherwise, the domain data Object model library and custom result encapsulation sub-module to customize data query;

Customized result encapsulation sub-module, which is used to encapsulate user-customized results; if this customization comes from the template library, set the identification bit of the template to indicate that it comes from an existing customized template, otherwise set the corresponding identification bit to indicate that it is a new customization; send to The request queue management sub-module of the data query server;

The data query server also includes a request queue management submodule, a request parsing submodule, a service proxy submodule, a service composition engine, a custom storage module, an execution submodule, a result generation submodule and an exception handling submodule;

The request queue management sub-module, the request queue management sub-module receives the data query request sent by the client and adds it to the request queue;

The request parsing sub-module takes out the first request in the request queue, and obtains the initial state, target state, constraints and result display format after parsing the request; if the request queue is not empty, judge whether the data query is a function in the template library If it is, the execution sub-module will directly obtain its service call sequence, otherwise it will call the service composition engine for service composition; if the request queue is empty, it will be sent to the exception handling sub-module for processing;

The service agent sub-module obtains the basic data query service information that can be used currently;

The service composition engine, according to the initial state and the target state, invokes the basic data that can be used currently to query service information, processes the query process in the initial state one by one, and performs service composition for each query process on the premise of satisfying the constraints of the intermediate process. Generate a service call sequence; if the service combination fails, it will be sent to the exception handling sub-module for processing;

Customize the storage module, store the corresponding function and service call sequence to the template library;

The execution sub-module calls the basic data query service in the basic data query service library according to the order of the service call process sequence, and coordinates the input and output relationship between services; if an exception occurs, it is sent to the exception handling sub-module for processing;

The result generation sub-module generates the final data query and analysis results according to the result display format described in the request, returns a response to the client, and displays the results;

The exception handling sub-module generates failure information and returns it to the client for display.

In the domain data object model library, the abstract classification of data is to extract the features of the data according to the idea of object-oriented; the attributes contained in the abstracted domain data object model are used as the attributes of the object model, and each domain data object The model is written in the form of client script as an editable control that can be customized to query the client, and describes its own data query requirements to the application system.

The domain data object model library includes logic controls, which use operations and logic operations to describe the relationship between two or more domain data object attributes, and each logic control is written in a script.

The basic data query service of the basic data query service library is in one-to-one correspondence with the domain data object models of the domain data object model library, providing atomic operations for domain data object query, and only through one or several attributes of the corresponding domain data object model , to obtain a collection of data object instances with this property from the background database.

Each basic data query service of the basic data query service library includes an atomic operation for querying associated attributes, an atomic operation for obtaining associated data objects and their attributes of corresponding domain data objects, and outputting associated data objects and attribute sets of associated data objects .

The customized data query performed by the customizable query client:

On the page of the customizable query client, select each editable control from the controls in the domain data model library, and specify each query content data object of this query; select the properties of each editable control as the result data list of this data query a row;

On the customization page, select each editable control from the visual controls in the domain data model library, and specify the conditional data object for this query; select the attribute of each editable control, select and fill in the attribute value for this attribute, and specify this time The query content data object for the selected query condition;

Arrange the order of the selected control properties, fill in the data list header of the custom page result, and specify the sorting column of the result list;

Name the newly customized data query, save this customization, and name the template corresponding to this customization.

The initial state obtained by the request parsing sub-module is composed of query process set, query result and service call sequence; the query process set includes the query process of the data object to be queried in this user request, and the query result and service call sequence are empty.

The target state obtained by the request parsing sub-module is composed of query process set, query result and service call sequence; the query process set is empty, and the query result is the target data object of each query process in the query process set, and the service call sequence is A free memory space.

The service composition engine extracts the source state of this service composition process, including the target data object of the query process, the query condition set and the empty service call sequence; adding the source state to the available state queue for temporarily storing data;

If the available status queue is empty, output service composition failure information, and the service composition process ends; otherwise:

Take the available state at the head of the queue from the available state queue as the current state;

If the query condition of the current state is empty, output the service call sequence of the current state, and the service combination process ends; otherwise, store the query condition of the current state in the temporary queue for storing the query condition;

If the temporary queue is empty, judge and process whether the available state queue is empty; otherwise:

Take the query condition at the head of the queue from the temporary queue as the current condition, search for the basic data query service that can be used for the current condition and find its corresponding atomic operation, if all the basic data query services have been called, then the available state queue Whether it is empty is judged and processed; otherwise:

Add the found atomic operation to the service call sequence of the current state in an equivalence relationship;

According to the current query condition of the current state, the successor state of the current state is generated, and the successor state is added to the available state queue, and whether the temporary queue is empty is judged and processed.

The service composition process of the service composition engine generates the successor state of the current state. If the domain data object corresponding to the current condition is the target object of the query process, the current condition is deleted from the current state, and the state after the deletion of the current condition is used as The subsequent state is added to the available state queue, and the generation is completed; otherwise, call the associated query atomic operation of the basic data query service corresponding to the current condition to obtain the set of associated attributes, add each associated attribute and the source of the attribute value, and delete the current state from the current state condition and add it to the available state queue as a successor state; then process the next associated attribute and add all the successor states to the available state queue until all associated attributes in the associated attribute set have generated successor states.

The invention provides a customizable query method and platform based on service combination, which has the advantages of:

(1) The present invention adopts the method that the user actively formulates the query purpose, query conditions and query result display format of a query, which can effectively deal with the user's diverse data query and analysis needs, allowing the user to modify the query process at any time according to their own needs, and No need to change the system structure.

(2) The present invention adopts the method of establishing basic data query services, based on abstract modeling of domain data, and uses object-oriented thinking to establish an independent basic data query service for each domain data object, which is in line with the logic of public thinking; Atomic operations are provided within the scope of the service, representing different input-output pairs, so that different query methods within the service can be clearly distinguished; the creation of atomic operations not only includes querying all domain data object instances with this attribute value by an attribute of the domain data object, And it includes querying all domain data object instances with these attribute values by multiple attributes that are usually used as joint attributes, avoiding multiple query operations on the background database and intersection of multiple data query results when only a single attribute query is provided operation to improve query efficiency.

(3) Construct data query processing logic based on service composition technology. Given the initial state and target state of the service composition process, it can dynamically generate service composition call sequences, and automatically respond to the variable data query requests of customized clients.

Description of drawings

Fig. 1 is a method flowchart of a customizable query method based on service composition in the present invention;

Fig. 2 is a platform structure diagram of a customizable query platform based on service combination in the present invention;

Fig. 3 is a flow chart of the customized data query function of a customizable query method based on service combination in the present invention;

FIG. 4 is a formalized description diagram of the user-customized result xml of a service-combination-based customizable query method in the present invention;

Fig. 5 is a service composition flowchart of a customizable query method based on service composition in the present invention;

FIG. 6 is a flow chart of generating a subsequent state of a customizable query method based on service composition in the present invention;

FIG. 7 is a user customization result in xml form in an embodiment of a service composition-based customizable query method of the present invention;

FIG. 8 is a state transition diagram of a service call sequence generation process in an embodiment of a service composition-based customizable query method of the present invention.

Detailed ways

The present invention proposes a customizable query method and platform based on service combination, which supports users to describe their own needs to the system, and the system dynamically assembles basic data query services according to user needs, and generates a combination of basic data query service call sequences that can meet user needs Services, execute combined services and display data query and analysis results according to user expectations.

A customizable query platform based on service composition, as shown in Figure 2, includes two parts: a customizable query client and a data query server.

The customizable data query client includes a template library, a domain data object model library, and a custom result encapsulation sub-module.

The template library is used to provide users with typical data query functions stored in the system. The template library is a collection of existing data query functions in the platform, which is derived from the user customizations read by the user customization storage sub-module and stored in the background database that have been successfully processed in the early stage of the system. The user selects a function in the template library, so that the page is displayed as a display of the query content, query conditions, and query result formats corresponding to this function.

The domain data object model library provides domain data object model visualization controls and logic controls.

The customized result encapsulation sub-module transforms the personalized query requirements customized by the user on the client side into a formal description.

The template library provides the existing data query function in the platform. When users query, they first check whether there is a data query function template that meets their own needs in the template library. If there is a template that meets the needs, select this template, modify the corresponding content in the template, and output user-defined From the result to the custom result encapsulation sub-module, the custom result encapsulation sub-module encapsulates the user-customized visualization result into an xml description, and sends it to the data query server as an HTTP request parameter; if there is no template that meets the requirements, it will turn to the domain data object model library , the user selects the query content, the corresponding domain data object control and its properties corresponding to the query conditions from the domain data object model library, and fills in the property values when necessary, specifies the display format of the data query analysis results, and outputs them to the customized result packaging sub-module. And still, the customized result encapsulation sub-module encapsulates the user-customized visualization result into an xml description, and sends it to the data query server as a parameter of the HTTP request.

The data query server includes a service proxy submodule, a request queue management submodule, a request parsing submodule, a service composition engine, an execution submodule, a result generation submodule, a custom storage submodule, an exception handling submodule and a basic data query service library. in:

The service agent sub-module is used to manage the basic services of domain data query.

Each basic data query service in the basic data query service library needs to register with the service agent sub-module, provide its own query information for a specific domain object model, and describe the domain provided by this service in the form of acceptable parameters All basic query operations of the data object model. The basic data query service can be added when necessary. It is necessary to ensure that the new basic data query service does not conflict with the original basic data query service, and is released to the service agent sub-module in a manner consistent with the release specification of the aforementioned basic data query service.

The request queue management sub-module is used to receive data query analysis requests sent by data query clients, and cache these requests.

The request parsing sub-module is used to parse the data query and analysis request, and convert the user-customized result carried in the request sent by the customizable query client into a data query request understandable by the service composition engine.

The service composition engine is used to combine the details of the available services provided by the service proxy sub-module, and according to the analysis results delivered by the request analysis module, generate service call sequences that can meet the requirements one by one for the query process included in the results.

The execution sub-module is used to sequentially execute the services in the sequence according to the service invocation sequence generated by the service composition engine.

The result generation sub-module is used to generate an HTTP response based on the result output by the execution sub-module according to the user-defined result format.

The customized storage sub-module is used to store the functions customized by the user this time, and the processing method of the customized functions.

The exception handling sub-module is used to return an error message to the data query client when the request parsing module cannot parse the data query request, the service composition engine cannot generate a service sequence that meets the user's needs, and the execution module fails to execute the service.

Each basic data query service in the basic data query service library provides its own information to the service agent sub-module, including the corresponding domain data object type and available atomic operations. The service agent sub-module maintains all the basic data query service information obtained by it into a list of currently available basic data query services for query by the service composition engine.

The request queue management sub-module of the data query processing server receives the data query request sent by the customized result encapsulation sub-module, and caches it into the request queue. The request parsing sub-module takes out the request at the head of the request queue, parses out the initial state and target state of the request, and sends the parsing result to the service composition engine. The service composition engine obtains the information of the currently available basic data query services provided by the service agent sub-module, generates a service call sequence capable of processing this data query request according to the initial state and target state of the request, and delivers the generated results to the execution sub-module module execution. The execution sub-module executes each basic data query service sequentially according to the service call sequence generated by the service composition engine, and outputs a set of data object instances containing user expected results to the result generation sub-module. The result generation sub-module extracts the content to be displayed specified in the user-customized result from the result output by the execution module, sorts the content to be displayed according to the sorting method specified by the user during customization, generates a response HTML page, and queries the client for customizable data Return an HTTP response. When the request parsing module cannot parse the data query request, the service composition engine cannot generate a service sequence that meets the user's needs, and the execution module fails to execute the service, the exception handling sub-module will capture these failure information, and the data query client will return an error message.

A customizable query method based on service composition, as shown in Figure 1, includes the following steps:

Step 1: domain analysis, abstract and classify the data in the domain, build editable controls that can customize the query client for the established model, store them in the domain data object model library, and build the corresponding basic data query service library. Each basic data query service in the basic data query service library corresponds to an object control in the domain data object model library. The basic data query service is a collection of multiple atomic operations. An atomic operation represents an input-output pair. The input is one or several attributes of the domain data object model, and the output is a collection of instances of the domain data object model with specified input attribute values. The atomic operation includes providing two methods for each attribute of the object control to query the domain data object instance according to the attribute, and to jointly obtain the data object instance according to several attributes at the same time. The combined attributes described in the latter are attributes that are taken into consideration during requirement analysis and are often combined as a query condition.

Step 2: Initialize the customizable query client, that is, initialize the domain data object model library of the customizable query client and the template library that stores existing customized functions. Controls in the domain data object model library are implemented as scripts. The templates in the existing customized function template library come from the data query and analysis requests that have been successfully processed in history; the template information is stored in the background database, and each template corresponds to a record in the database, which includes template ID, template name, and template corresponding There are four parts: the configuration information of the template (that is, the query content and query conditions corresponding to this template), and the service call sequence corresponding to the template function. When the user opens the customized page, the display form of the domain object model library is a collection of editable controls, and the display form of the existing customized function template library is a collection of existing customized function template names.

The user compares his personalized needs with the existing functions in the template library. If there is a function that meets the user's needs this time, select a specific template name from the template library, read the template ID corresponding to the template name and the template configuration information. The template corresponds to the customized query content, query conditions, and result display format, and generates editable controls and properties corresponding to the query content, editable controls and properties corresponding to the query conditions, and the header of the result display list. The user constructs a new query request with the same function by adjusting the display format of the result and refilling the value of the customized query condition attribute.

If there is no function that meets the needs of the user, the customizable query client can customize the data query and customize a new data query function, as shown in Figure 3, including the following steps:

A) Customize the query content. On the page of the customizable query client, select each editable control from the domain data model library control, drag and drop it into the query content customization box to specify each query content data object of this query, and the query result data comes from the query content data object Some attributes, check the attributes of each editable control as a column of the result data list of this data query, that is, indicate which attributes of the query content data object the query result data is.

B) Customize query conditions. On the customization page, select each editable control from the visual control of the domain data model library, drag and drop it into the query condition customization box to specify the condition data object of this query, check the properties of these editable controls and select, Fill in the attribute value, that is, specify which attribute of the data object the parameter used as the query condition comes from, and determine the attribute value of this attribute. Specify which query content data object the selected query condition is used for, and use an arrow to point from the query condition data object control to the query content data object control to indicate association.

C) Customize the display format of the results. Arrange the order of the control attributes checked in A) according to user preference, fill in the data list header of the custom page result, and specify the sorting column of the result list, and put the sorting column in the first column.

D) Name the new custom data query function. The user checks the button of saving this customization, and enters the name of the template corresponding to the function customized this time in the template name text box.

Step 3: After the user customization process is completed, the customized result packaging sub-module of the customizable query client encapsulates the user customized results. If this customization comes from the template library, set the corresponding flag to indicate that it comes from an existing custom template, if not, then this customization is a brand new customization, and set the corresponding flag to indicate that it is a new customization. The user-customized result is packaged as an xml file, as shown in Figure 4.

With request as the root node, it consists of several processes and a resultForm. process represents a query process, which consists of a goal node (that is, the query content) and a parameter (that is, the query condition) node. The goal node is composed of a data object node model, and the model node includes a name node of the model and multiple property nodes of the model. The property node contains the name of the property and the value of the property (empty here). The parameter node is composed of multiple model nodes. Each model node includes a name node of the model, and multiple property nodes of the model. The property node contains the name of the property and the value of the property, and the value is not empty. resultForm represents the result format, which contains several column nodes. Each column node contains the column name (that is, the header name of the data query result list) name, and the value of the column (that is, the data source of this column data) value. The encapsulation result is used as a parameter, and the data query http request is sent to the data query processing server.

Step 4: The request queue management sub-module of the data query server receives the data query request sent by the client and adds it to the request queue.

Step 5: The request parsing sub-module takes out the first request in the request queue, and obtains the initial state, target state, constraint conditions and result display format after parsing the request. The initial state consists of three parts: the set of query procedures currently to be executed, the query results, and the sequence of service calls. The set of query processes to be executed comes from the content of the process node under the request node in the parameter xml file carried in this data query request. Since the process of generating the composite service has not yet been executed, the query result is empty and the sequence of service calls is empty. The target state is composed of three parts: the set of currently to-be-executed query procedures, the query results, and the sequence of service calls. In the target state, all query processes have been processed, and the set of query processes to be executed is empty. The query result indicates that each query process has been successfully executed, and the target data objects of each query process can be successfully queried, so the query result is a collection of target data objects of each query process. Service call sequence Because the service composition process that generates the service call sequence has not been executed, it will be filled after the service composition process is completed, and it is currently a free memory space. If the request queue is not empty, it is judged that the data query is from a function in the custom function template library. If so, the background database is directly queried according to the template name of the source to obtain the data query function corresponding to this template. service call sequence, go to step 8; otherwise, call the service composition engine to perform service composition, as shown in step 6.

If the queue is found to be empty, that is, the request parsing fails, go to step ten.

Step 6: According to the initial state and the target state, the service composition engine sub-module calls the service proxy sub-module to obtain the currently available basic data query service information, and handles the query process in the initial state one by one. Each query process performs service composition to generate a service invocation sequence.

Generate a service call sequence, as shown in Figure 5, including the following steps:

Step1: Extract the source state So of this service composition process from the query process, So includes the target data object of this query process, query condition set, and empty service call sequence. The target data object indicates the domain data object type of the data that needs to be obtained in this query process; the query condition set is the query condition selected and filled in by the user on the customized page, and is the condition associated with the target data object of the query process; the service call in the initial state The sequence is empty. Add So to the available state queue TmpS for temporary storage of data.

Step2: Determine whether TmpS is empty, if it is empty, it means that there is no state available for processing, then output service combination failure information, and the service combination process ends; if it is not empty, go to step3.

Step3: Take the available state at the head of the queue from TmpS as the current state S, and the next available state in TmpS is placed at the head of the queue.

Step4: Determine whether the query condition of the current state S is empty, and if it is empty, output the service call sequence of the current state S, and the service combination process ends; if it is not empty, store the query condition of S in the queue A that temporarily stores the query condition.

Step5: If A is empty, it means that in the current state, all query conditions are used respectively and the corresponding successor state is generated, go to step2; otherwise, there are still unused query conditions and go to step6.

Step6: Take the query condition at the head of the queue from A as the current condition, and the next query condition in A is placed at the head of the queue. Search for the basic data query services that can be used for the current conditions and find their corresponding atomic operations. If all basic data query services have been called, it indicates that there are query conditions that cannot be processed in this state, so the state cannot reach the target state, then Give up the current state and go to step2; otherwise, go to step7.

Step7: Add the found atomic operation to the service call sequence of S. If multiple atomic operations of a basic data query service can be searched according to the current conditions, choose among the atomic operations with the same input to have multiple conditional inputs. Atomic operations of different inputs must be called. These atomic operations are added to the service call sequence of the current state with an equivalence relationship. Each atomic operation is executed separately, and then the intersection of the execution results is taken as the output of the basic data query service call.

Step8: According to the current query condition of the current state, generate the successor state of the current state, and add the successor state to the available state queue.

As shown in Figure 6, the generation of the subsequent state includes two situations: the first is that the domain data object corresponding to the current condition is the target object of this query process, then the current condition is deleted from the current state, indicating that the current condition has been used as the basic data The input condition of the atomic operation of the query service cannot be used by the basic data query service in the future. The state after deleting the current condition is used as the successor state, and TmpS is added to indicate that a new usable state has been generated, and the generation is complete; the domain data object corresponding to the second current condition is not the target object of the query process, in which case a successor state is generated The method is to call the associated query atomic operation of the basic data query service corresponding to the current condition to obtain the associated attribute set. The attribute value source of the associated attribute is the output of the atomic operation corresponding to the basic data query service of the current condition. Delete the current condition from the current state and add each associated attribute and attribute value source separately, indicating that the current condition cannot be directly used to filter the target data object of this query process, but the current condition can generate other conditions through the association relationship, and these generated conditions It can be used to filter the target data object of this query process. Add the newly generated state to TmpS as a successor state, indicating that a new usable state has been generated. Process the next associated attribute until all associated attributes in the associated attribute set have generated a successor state, and after all the successor states are added to TmpS, the generation is completed.

After the successor state is generated, continue to process the next condition and go to step5.

The service call sequences generated by multiple query processes can be processed in parallel, that is, the query processes are independent of each other, and the service call sequences generated by two or more query processes can be executed simultaneously.

If the service combination fails and the service invocation sequence cannot be generated, go to step ten.

Step 7: The custom storage module stores the corresponding function and service call sequence into the template library, and becomes a template information record. The template ID of the record is automatically generated by the background database; the template name is the template name in the user-customized result xml; the template corresponds to The query content in the configuration information is the query content in the user-customized result xml, and the query condition is the query condition in the user-customized result xml. The query condition only includes the data object and attribute to which the query condition belongs, excluding the attribute value; the service call corresponding to the template The sequence is the service call sequence generated in step 6 above, where the input of each basic data query service atomic operation is empty.

Step 8: The execution sub-module invokes the basic data query service in the basic data query service library according to the order of the service calling process sequence, and generates intermediate results and final results. The execution sub-module coordinates the input and output relationship between services. When reading the service call sequence, if the atomic operation input of a certain basic data query service is not the parameter directly input by the user when customizing, but the service call sequence has been executed For the atomic operation of the basic data query service, the intermediate result generated by the executed atomic operation of the basic data query service is used as the input of the basic data query service. On the contrary, the parameters entered by the user during customization are directly used as the input of the atomic operation of the basic data query service, and the atomic operation is executed. After the entire service call sequence is executed, the final result is output. If an exception occurs and the execution fails, go to step ten.

Step 9: The result generation sub-module generates the final data query and analysis result according to the result display format described in the request, and returns a response to the client in the form of a browser page; and displays the result.

Step 10: The exception handling sub-module generates failure information and returns it to the client for display.

The realization premise of a customizable query method based on service combination in the present invention is the abstract modeling of domain data, classifying the data in the domain, and constructing the domain data object model library. Then build a data query service according to the basic query requirements of each domain data object, and publish it as a basic service to the service proxy module of the data query server.

The present invention will be described in further detail below in conjunction with the commodity inquiry of accompanying drawing and an online shop system as embodiment:

In this embodiment, there is an initial query request from the user, and the query product is HP brand, which was launched after 2007, and currently has computer data in stock. It is required to display the computer model, quotation, and configuration information, sorted according to the order of quotation from low to high . It is assumed that in this embodiment, there is no template with a similar function available for direct calling in the template library. The processing steps are as follows:

Step 1: domain analysis, the computer is the computer data object has attributes: product number, model, brand, manufacturer, time to market, quotation, configuration information; the inventory product is the storage data object has attributes: product number, inventory, cost price , selling price, administrator information. The relationship between the computer data object and the storage data object is as follows: the item number attribute of the computer data object is equal to the item number attribute of the storage data object. The corresponding basic data query services are computerQservice and storageQService. computerQService provides the atomic operation computerOperByBrand for obtaining computer object instances according to the computer brand and computerOperByTTM, the atomic operation computerOperByTTM for obtaining computer object instances according to the computer brand and price, and computerOperByBrandAndPrice for obtaining computer object instances according to the computer product number. The atomic operation computerOperByID of the object instance, the atomic operation computerOperRela of the associated query; storageQService provides the atomic operation storageOperByStock and the atomic operation storageOperRela of the associated query to obtain the object instance of the stock commodity according to the inventory.

Step 2: The user enters the customizable data query client page. At this time, the data query client initializes to obtain the domain data object model library, which includes two domain data object controls computer and storage, implemented in JavaScript in this embodiment. The content of the template library is empty; there is no data query function corresponding to the query computer in the template library, and the user needs to start a new function customization process.

A) Customize the query content. The user drags the computer control from the customization page to the query content customization box, and checks the model, quotation, and configuration information as the result attributes;

B) Customize query conditions. The user drags the computer control from the customization page to the query condition customization box, checks the brand attribute, fills in the attribute value as "HP", checks the time-to-market attribute, and fills in the attribute value as ">2007"; drags and drops the inventory from the customization page Add the product control to the query condition customization box, check the inventory attribute, and fill in the attribute value as ">0". Use arrows to establish the relationship between query conditions and query content: computer: brand=HP; TTM=>2007 and storage: stock>0, both of which are associated with the query content computer;

C) Customize the display format of the results. Customize the result of this request to display the header of the data list as follows:

Table 1 Customized result display data list header

show Valence chromium Computer Model Computer Configuration value computer: price (sort column) computer: model computer: configure

computer:model indicates the computer model, computer:price indicates the computer price, computer:configure indicates the configuration information of the computer, and the results are sorted according to the first column attribute, that is, the price column.

D) Check the button to save this customization, and enter the name of the template corresponding to this customization as "Purchasable Computer Product Query".

In this embodiment, there is no user-customized function in the template library. If there is a template named "Query Computer Products" with user-customized functions, the processing process is as follows: select this template, and the page query content part is displayed as a computer data object, which has been checked The attributes of are model, quotation, and configuration information. The query condition part of the page is displayed as a computer data object and a storage data object. The checked attributes of the computer data object are launch time and brand, and the attribute value is empty. The checked attributes of the storage data object are inventory, and the attribute value is empty. The query result format is empty and filled by the user. The user fills in the attribute value of ">2007" for the time-to-market attribute of the computer data object used as the query condition, "HP" for the brand attribute attribute value, and ">0" for the inventory attribute of the storage data object used as the query condition ”; fill in the result format as shown in Table 1 above.

Step 3: Client customized result encapsulation The sub-module encapsulates the user customized result, as shown in Figure 7, expressing the user customized result in the form of xml. With request as the root node, it consists of a process and a resultForm. The process consists of a goal node and a parameter node. The goal node is composed of a data object node model. The name of the model node is computer, indicating that the encapsulated data object is a computer. The model has three property nodes. The names of the property nodes are model, price, and configure, which respectively represent the model, price, and configuration information of the computer. The parameter node is composed of multiple model nodes. The name of the first model node is computer, the attribute property node, the name of the property node is brand, and the value is HP. The rest of the model is similar to this model. resultForm represents the result format, which contains several column nodes. The first column node contains the column name price, and the column value computer:price. The rest of the columns are similar to this column. Use this xml as a parameter to send a data query http request to the data query processing server.

Step 4: The request queue management sub-module of the data query processing server receives the data query request sent by the client, and stores the request in the request queue.

Step 5: The request parsing module uses DOM to parse the first data query and analysis request in the request queue. In the initial state, PROCESS is used to represent the set of query processes, and target=(computer) means that the target data object of this query process is computer, parameter =((computer:brand=HP; TTM=>2007)(storage:stock>0)) indicates that the query condition of this query process is that the brand of the computer is HP, the TTM of the time to market is after 2007, and the stock quantity of the stock commodity object is If it is greater than zero, there is inventory; RESULT indicates the query result, and PLAN indicates the service call sequence. The result of parsing the preceding request is:

Table 2 request parsing results

initial state {PROCESS(target=(computer)parameter=((computer: brand=HP; TTM=>2007)(storage: stock=>0))) RESULT()

plan()} target state {RESULT(computer)PLAN(planForPROCESS)} Restrictions none

In the target state, the query process PROCESS has been processed and is empty, the query result RESULT is the set of target objects in the successfully processed PROCESS, that is, computer, and the service call sequence PLAN is planForPROCESS, which refers to the service call generated by the PROCESS in the initial state sequence.

If the request queue is not empty, it is judged that the query is not a type of function in the template library, and a service call sequence needs to be generated by the service composition engine sub-module. If it is one of the functions in the template library, query the background database according to the template name, and obtain the query request processing service call sequence corresponding to the template by the execution sub-module, and use the parameter in the initial state as the parameter value to match the basic data query service atom Operation, as the input of the atomic operation, go to step 8. If the request queue is empty, go to step ten.

Step 6: The service composition engine generates a service call sequence for this data query request. This data query request contains a PROCESS, as shown in Figure 8, and the process of generating a service call sequence is detailed as follows:

PROCESS: The query content is computer, the parameters are (computer: brand=HP; TTM=>2007) and (storage: stock>0), the initial state of this query process: (target=(computer)parameter=((computer: brand= HP; TTM=>2007)(storage:stock>0))plan=()), the target status of this query process: (target=(computer)parameter=()plan=(*)), * means pre-allocated idle space.

Step1: Extract the source state S ₀ of this service composition process from the query process, S ₀ =(target=(computer)parameter=((computer: brand=HP; TTM=>2007)(storage: stock>0)) plan=()); Add S ₀ to the available state queue TmpS for temporary storage of data, TmpS={S ₀ }.

Step2: Determine that TmpS is not empty, go to step3.

Step3: Take So at the head of the queue from TmpS as the current state S.

step4: The query condition of S is that parameter=((computer: brand=HP; TTM=>2007)(storage:stock>0)) is not empty, put this condition in the queue A of the temporary storage query condition, that is, A={ (computer: brand=HP; TTM=>2007), (storage: stock>0)}.

step5: A is not empty, go to step6.

step6: Take out the query condition at the head of the queue from A (computer: brand=HP; TTM=>2007) as the current condition, search for the basic data query service that can be used for the current condition is computerQService, and its corresponding atomic operations are computerOperByBrand and computerOperByTTM, The atomic operation computerOperByBrandAndPrice has brand as an input parameter, but the other input parameter is price, and the current condition does not contain price, so this atomic operation is not available for the current condition. Go to step7.

Step7: Add (computerQService: computerOperByBrand(HP))(computerQService:computerOperByTTM(>2007)) into the service call sequence plan of S, at this time plan=(computerQService:computerOperByBrand(HP))∧(computerQService:computerOperByTTM(>2007 )).

step8: Generate a new state S'=(target=(computer)parameter=((storage:stock>0))plan=((computerQService:computerOperByBrand(HP))∧(computerQService:computerOperByTTM(>2007)))), current The domain data object corresponding to the condition is computer, and the target object is also computer, so directly add S' to the available TmpS, at this time TmpS={(target=(computer)parameter=((storage:stock>0))plan=((computerQService :computerOperByBrand(HP))∧(computerQService:computerOperByTTM(>2007))))}; turn to step5a.

step5a: A={(storage: stock>0)} is not empty, go to step6a.

Step6a: Take out the query condition (storage: stock>0) at the head of the queue from A as the current condition, search for the basic data query service that can be used for the current condition is storageQService, and its corresponding atomic operation is storageOperByStock, go to step7a.

Step7a: Add storageQService:storageOperByStock(>0) into the service call sequence plan of S, at this time plan=(storageQService:storageOperByStock(>0)).

step8a: Generate a new state S'=(target=(computer)parameter=((computer: brand=HP; TTM=>2007)) plan=(storageQService:storageOperByStock(>0))), the domain data object corresponding to the current condition Because storage is inconsistent with the target field data object computer, call storageQService’s associated query basic operation storageOperRela to get the result storage: ID equivalent to computer: ID, add the associated attribute to the query condition of S’ to generate a subsequent state (target=(computer)parameter= ((computer: brand=HP; TTM=>2007; ID=(storageQService:storageOperByStock(>0))))plan=(storageQService:storageOperByStock(>0))), and add subsequent state to TmpS, TmpS= {(target=(computer)parameter=((storage:stock>0))plan=((computerQService:computerOperByBrand(HP))∧(computerQService:computerOperByTTM(>2007)))), (target=(computer)parameter= ((computer: brand=HP; TTM=>2007; ID=(storageQService:storageOperByStock(>0))))plan=(storageQService:storageOperByStock(>0)))}; turn to step5b.

step5b: A is empty, go to step2c.

step2c: TmpS is not empty, go to step3c.

step3c: Take out the state at the head of the queue from TmpS as the current state S, S={(target=(computer)parameter=((storage:stock>0))plan=((computerQService:computerOperByBrand(HP))∧(computerQService :computerOperByTTM(>2007))))}.

step4c: The query condition of the current state S is not empty, store the query condition of S in A, then A={(storage: stock>0)}.

step5c: A is not empty, turn to step6c.

Step6c: Take out the query condition (storage: stock>0) at the head of the queue from A as the current condition, search for the basic data query service that can be used for the current condition is storageQService, and its corresponding atomic operation is storageOperByStock, go to step7c.

step7c: add storageQService:storageOperByStock(>0) into the service call sequence plan of S, then plan=(computerQService:computerOperByBrand(HP))∧(computerQService:computerOperByTTM(>2007)), storageQService:storageOperByStock(>0) .

step8c: Generate a new state S'=(target=(computer)parameter=() plan=((computerQService: computerOperByBrand(HP))∧(computerQService: computerOperByTTM(>2007)), (storageQService: storageOperByStock(>0))) , the domain data object corresponding to the current condition is storage and the target domain data object computer is inconsistent, call storageQService’s associated query basic operation storageOperRela, get the result storage: ID is equivalent to computer: ID, add the associated attribute to the query condition of S’ to generate a subsequent state (target=(computer)parameter=((computer: ID=(storageQService:storageOperByStock(>0))))plan=((computerQService:computerOperByBrand(HP))∧(computerQService:computerOperByTTM(>2007)),storageQService:storageOperByStock (>0))), and add subsequent state to TmpS, TmpS={(target=(computer)parameter=((computer: brand=HP; TTM=>2007; ID=(storageQService:storageOperByStock(>0) ))) plan=(storageQService:storageOperByStock(>0))), (target=(computer)parameter=((computer:ID=(storageQService:storageOperByStock(>0))))plan=((computerQService:computerOperByBrand(HP ))∧(computerQService: computerOperByTTM(>2007)), storageQService: storageOperByStock(>0)))}, turn to step5d.

step5d: A is empty, turn to steD2e.

step2e: TmpS is not empty, go to step3e.

step3e: Take out the state at the head of the queue from TmpS as the current state S, S={(target=(computer)parameter=((computer: brand=HP; TTM=>2007; ID=(storageQService:storageOperByStock(>0) )))plan=(storageQService:storageOperByStock(>0)))}.

step4e: The query condition of the current state S is not empty, store the query condition of S in A, A={(computer: brand=HP; TTM=>2007; ID=(storageQService:storageOperByStock(>0)))};

step5e: A is not empty, go to step6e.

step6e: Take the query condition at the head of the queue from A (computer: brand=HP; TTM=>2007; ID=(storageQService:storageOperByStock(>0)))) as the current condition, and search for the basic data query service that can be used for the current condition It is computerQService, and its corresponding atomic operations are computerOperByBrand, computerOperByTTM and computerOperByID, turn to step7e.

Step7e: Add (computerQService: computerOperByBrand(HP))∧(computerQService: computerOperByTTM(>2007))∧computerQService:computerOperByID(storageQService:storageOperByStock(>0)) into the service call sequence plan of S, then plan=(storageQService :storageOperByStock(>0)), (computerQService:computerOperByBrand(HP))∧(computerQService:computerOperByTTM(>2007))∧(computerQService:computerOperByID(storageQService:storageOperByStock(>0))).

step8e: Generate a new state S'=(target=(computer)parameter=() plan=(storageQService:storageOperByStock(>0)), (computerQService:computerOperByBrand(HP))∧(computerQService:computerOperByTTM(>2007))∧( computerQService: computerOperByID(storageQService: storageOperByStock(>0)))), the domain data object corresponding to the current condition is computer, and the target object is also computer, directly add S' to the available TmpS, at this time TmpS={(target=(computer) parameter=((computer: ID=(storageQService:storageOperByStock(>0))))plan=((computerQService:computerOperByBrand(HP))∧(computerQService:computerOperByTTM(>2007)),storageQService:storageOperByStock(>0))) , (target=(computer)parameter=() plan=((storageQService:storageOperByStock(>0)), (computerQService:computerOperByBrand(HP))∧(computerQService:computerOperByTTM(>2007))∧computerQService:computerOperByID(storageQService:storageOperByStock (>0)))}, turn to step5f.

step5f: A is empty, turn to step2g.

step2g: TmpS is not empty, go to step3g.

step3g: Take out the state at the head of the queue from TmpS as the current state S, S={(target=(computer)parameter=((computer: ID=(storageQService:storageOperByStock(>0))))plan=((computerQService: computerOperByBrand(HP))∧(computerQService: computerOperByTTM(>2007)), storageQService: storageOperByStock(>0)))}.

step4g: Determine that the query condition of the current state S is not empty, store the query condition of S in the queue A for temporarily storing the query condition, A={(computer: ID=(storageQService:storageOperByStock(>0)))}.

step5g: A is not idling step6g.

step6g: Take the query condition at the head of the queue from A (computer: ID=(storageQService:storageOperByStock(>0))) as the current condition, and search for the basic data query service that can be used for the current condition as computerQService, which has been called by the current state S However, give up the current state S and go to step2h.

step2h: TmpS is not empty, go to step3h.

step3h: Take out the state at the head of the queue from TmpS as the current state S, S=(target=(computer)parameter=() plan=((storageQService:storageOperByStock(>0)), (computerQService:computerOperByBrand(HP))∧ (computerQService: computerOperByTTM(>2007))∧computerQService: computerOperByID(storageQService:storageOperByStock(>0)).

step4h: The current status query condition is empty, and the output service call sequence: (storageQService: storageOperByStock(>0)), (computerQService: computerOperByBrand(HP))∧(computerQService: computerOperByTTM(>2007))∧(computerQService: computerOperByID(storageQService: storageOperByStock(>0)))), the service composition process of the query process is completed. Since this query request only includes this query process, the service combination process of this query request is completed.

If service composition fails, go to step ten.

Step 7: The custom storage module stores the successfully generated service call sequence and the template corresponding to the data query request function in the template library. In the stored template records, the template ID is automatically generated by the background database, the template name is the template name in the user-defined package result xml, that is, "purchasable computer product query", and the corresponding configuration information of the template is parsed from the user-customized package result xml The query content corresponding to this data query is computer, and the query condition is ((computer: brand = "; TTM = ") (storage: stock = ")), indicating that the query condition is the brand attribute, time-to-market attribute, and inventory of the computer object The inventory attribute of the commodity object. The service call sequence corresponding to the template is the service call sequence generated by the aforementioned service combination process, but does not include the parameter value, namely (storageQService:storageOperByStock()), (computerQService:computerOperByBrand())∧(computerQService: computerOperByTTM())∧(computerQService:computerOperByID(storageQService:storageOperByStock()))), which means first calling the storageOperByStock atomic operation of the storageQService basic data query service, which means querying all instances of the stock commodity object according to the stock quantity. Then call computerQService separately The computerOperByBrand atomic operation of the basic data query service (represents querying computer products by brand), computerOperByTTM atomic operation (represents querying computer products by launch time), computerOperByID atomic operation (represents querying computer products by product number), and the above three atomic operations Take the intersection of the output results to get the final result of this service call sequence.

Step 8: The execution sub-module invokes the basic services in sequence according to the generated service call sequence, uses the input specified in the sequence as the input of the atomic operation of the basic data query service, and finally outputs a set of computer data object instances. computerQService: computerOperByID() The input of this atomic operation is the ID attribute of the storage object that is the output result of the previous execution of storageQService: storageOperByStock(>0). When the atomic operation computerQService: computerOperByID() is executed, take out storageQService: storageOperByStock(>0) As a result, the ID attribute of each storage instance in the storage instance collection forms a list, which is used as the input of computerQService:computerOperByID().

If the execution fails due to service call failure, wrong parameter type, etc. during execution, go to step 10.

Step 9: According to the result format described in step 2, use the result data object instance output in step 8 as the data source, extract the price, model, and configuration data from the computer data object of the data source, and generate the data query result of this time. All row data is sorted according to the computer:price column, and the server-side response HTML page is generated with the sorted data list as the data source, and the response is returned.

Step 10: The exception handling sub-module returns failure information to the client, indicating the failure type and detailed information.

Claims (20)

1. the customizable querying method based on Services Combination is characterized in that, is made up of following steps:

Step 1: set up customizable inquiring client terminal and data query server, and data in the field are carried out abstract classification, for but the field data object model that takes out makes up edit control, deposit the data object model storehouse, field in the customizable inquiring client terminal in, and in the data query server, make up corresponding master data inquiry service, deposit master data inquiry service storehouse in;

Step 2: the data object model storehouse, field of the customizable inquiring client terminal of initialization and the template base that is used to store existing customization function of customizable inquiring client terminal; The user compares self individual demand with the existing function in the template base, satisfy user's template of the function of demand this time if exist, from template base, select the template of specific function, the page is showed the query contents, querying condition of the corresponding customization of template for this reason, display format as a result, and adjust display format as a result, rewrite the querying condition property value of customization, the new request of structure identical function; Otherwise the customization result by data object model storehouse, field and customizable inquiring client terminal encapsulates the inquiry of submodule customization data;

Step 3: the customization result of customizable inquiring client terminal encapsulates submodule encapsulation customization result; If this customization derives from template base, set the flag of template and represent to derive from existing custom built forms, otherwise the flag of setting template is expressed as new customization; The customization result encapsulate submodule encapsulated result as parameter, send the request queue management submodule of data query request to the data query server;

Step 4: request queue management submodule receives the data query request that customizable inquiring client terminal sends, and joins in the request queue;

Step 5: the request analysis submodule of data query server takes out first request in the request queue, resolves this request back acquisition original state, dbjective state, constraint condition and display format as a result; If request queue is not empty, judge whether this data query is a kind of of function in the template base, if, then directly obtain the service call sequence of template in this data query corresponding templates storehouse by the implementation sub-module of data query server, change step 8, otherwise the Services Combination engine that calls the data query server carries out Services Combination; If request queue is empty, then change step 10;

Step 6: original state and dbjective state that the Services Combination engine obtains according to the request analysis submodule, call the current master data inquiry service information that can use of service broker's submodule acquisition of data query server, handle the query script in the original state one by one, under the prerequisite that satisfies pilot process constraint condition, for each query script carries out Services Combination, generate the service call sequence; If the Services Combination failure then forwards step 10 to;

Step 7: by the corresponding function of the customization memory module of data query server storage and service call sequence to template base;

Step 8: implementation sub-module is according to the master data inquiry service in the master data inquiry service storehouse of the sequence call data query server of service call sequence, and input/output relation between coordination service, treat that whole service call sequence is finished, the output net result; If unusually then forward step 10 to;

Step 9: generate submodule by the result of data query server the step 8 net result is generated final data query analysis result according to the described display format as a result of request,, carry out the result and show to customizable inquiring client terminal echo reply with the form of browser page;

Step 10: the abnormality processing submodule by the data query server generates failure information, and returns to customizable inquiring client terminal demonstration.

2. according to the described a kind of customizable querying method of claim 1, it is characterized in that based on Services Combination: in the described step 1 data in the field are carried out abstract classification is according to OO thought, data are carried out feature extraction; The attribute that the field data object model that takes out is comprised is as the attribute of field data object model, but each field data object model is written as the edit control of customizable inquiring client terminal in the mode of client script, describes the data query demand of self to the data query server.

3. according to the described a kind of customizable querying method of claim 2 based on Services Combination, it is characterized in that: the data object model storehouse, field in the described step 1 comprises logical control, with computing and logical operation the incidence relation between the data object attribute of two or more fields is described, each logical control is write with script mode.

4. according to the described a kind of customizable querying method of claim 1 based on Services Combination, it is characterized in that: the master data inquiry service in the master data inquiry service storehouse in the described step 1 is corresponding one by one with the field data object model in data object model storehouse, field, the atomic operation of field data object inquiry is provided, only by one or several attribute of corresponding field data object model, acquisition has the set of the data object example of this attribute from data query data in server storehouse.

5. according to the described a kind of customizable querying method of claim 4 based on Services Combination, it is characterized in that: each master data inquiry service in the master data inquiry service storehouse in the described step 1 comprises the atomic operation of an inquiry relating attribute, inquiry obtains the associated data object and the attribute thereof of corresponding field data object, the community set of output associated data object and associated data object.

6. according to the described a kind of customizable querying method based on Services Combination of claim 1, it is characterized in that: the customization data inquiry in the described step 2 comprises the steps:

Step a: but each edit control on the page of customizable inquiring client terminal, from the control in domain data model storehouse, selected, specify each query contents data object of this inquiry; But select the row of the attribute of each edit control as the result data tabulation of this data query;

Step b: but each edit control on customized web page, from the visual control in domain data model storehouse, selected, specify the condition data object of this inquiry; But property value is selected, filled in to the attribute of each edit control of selection also for this attribute, and the query contents data object of specifying this querying condition of selecting to be used for;

Step c: arrange the sequencing of the control property of selecting, fill customized web page result's data list gauge outfit, and the row as sort key of designated result tabulation;

Steps d: the data query of the new customization of name, preserve this customization, and name the corresponding template of this customization.

7. according to the described a kind of customizable querying method of claim 1 based on Services Combination, it is characterized in that: the original state in the described step 5, form by query script set, Query Result and service calling sequence, wherein, the query script set comprises that this user asks the query script of each data object correspondence of needs inquiry, and Query Result and service calling sequence are empty.

8. according to the described a kind of customizable querying method of claim 1 based on Services Combination, it is characterized in that: the dbjective state in the described step 5, form by query script set, Query Result and service calling sequence, wherein, the query script set is for empty, Query Result is the target data object of each query script in the query script set, and the service call sequence is the memory headroom of one period free time.

9. according to the described a kind of customizable querying method based on Services Combination of claim 1, it is characterized in that: the Services Combination in the described step 6 comprises the steps:

Step1: the Services Combination engine extracts the source state of this Services Combination process, comprises target data object, querying condition set and the empty service call sequence of query script; Interpolation source state is to the upstate formation that is used for temporary storaging data;

Step2: if the upstate formation is empty, then export the Services Combination failure information, the Services Combination process finishes; Otherwise change step3;

Step3: taking-up is positioned at the upstate of queue heads as current state from the upstate formation;

Step4: if the querying condition of current state is empty, then export the service call sequence of current state, the Services Combination process finishes; Otherwise the querying condition of current state is stored in the interim formation of storing queries condition;

Step5:, change step2 if interim formation is empty; Otherwise change step6;

Step6: from interim formation, take out the querying condition conduct that is positioned at queue heads and work as precondition, search can be used for when the master data inquiry service of precondition and search its corresponding atomic operation, used if all master data inquiry services are all modulated, then changeed step2; Otherwise change step7;

Step7: the service call sequence of the atomic operation that finds being added to current state with relation of equivalence;

Step8: at the current querying condition of current state, generate the follow-up state of current state, and follow-up state is added the upstate formation, change step5.

10. according to the described a kind of customizable querying method of claim 9 based on Services Combination, it is characterized in that: when described step8 generates the follow-up state of current state, if when the field data object of precondition correspondence is the destination object of this query script, then precondition is worked as in deletion from current state, and the state that will delete behind precondition adds the upstate formation as follow-up state, and generation finishes; Otherwise, call correlation inquiry atomic operation when the master data inquiry service of precondition correspondence, each relating attribute and property value source is added in the set of acquisition relating attribute respectively, and deletion adds the upstate formation when precondition and as follow-up state from current state; Then handle next relating attribute and all follow-up states are added the upstate formation, all relating attributes have all generated follow-up state in the relating attribute set.

11. the customizable searching platform based on Services Combination is characterized in that, comprises customizable inquiring client terminal and data query server; Customizable inquiring client terminal comprises data object model storehouse, field, is used for data in the field are carried out abstract classification, but deposits the edit control that makes up for the field data object model that takes out in; The data query server comprises master data inquiry service storehouse, but the master data inquiry service of structure and storage edit control correspondence;

Customizable inquiring client terminal comprises that also template base, customization result encapsulate submodule;

Template base, be used for storing existing customization function, after the initialization template base, the user compares self individual demand with the existing function in the template base, satisfy user's template of the function of demand this time if exist, and selects the template of specific function from template base, the page is showed the query contents, querying condition of the corresponding customization of template for this reason, display format as a result, and adjust display format as a result, rewrite the querying condition property value of customization, the new request of structure identical function; Otherwise encapsulate the inquiry of submodule customization data by data object model storehouse, field and customization result;

The customization result encapsulates submodule, is used for encapsulating the customization result; If this customization derives from template base, set the flag of template and represent to derive from existing custom built forms, otherwise the flag of setting template is expressed as new customization; The customization result encapsulate submodule encapsulated result as parameter, send the request queue management submodule of data query request to the data query server;

The data query server comprises that also request queue management submodule, request analysis submodule, service broker's submodule, Services Combination engine, customization memory module, implementation sub-module, result generate submodule and abnormality processing submodule;

Request queue management submodule, request queue management submodule receives the data query request that customizable inquiring client terminal sends, and joins in the request queue;

The request analysis submodule takes out first request in the request queue, resolves this request back acquisition original state, dbjective state, constraint condition and display format as a result; If request queue is not empty, judge whether this data query is a kind of of function in the template base, if, then directly obtain the service call sequence of template in this data query corresponding templates storehouse by implementation sub-module, carry out Services Combination otherwise call the Services Combination engine; If request queue is empty, then sends to the abnormality processing submodule and handle;

Service broker's submodule obtains the current master data inquiry service information that can use;

The Services Combination engine, the original state and the dbjective state that obtain according to the request analysis submodule, call the current master data inquiry service information that can use, handle the query script in the original state one by one, under the prerequisite that satisfies pilot process constraint condition, for each query script carries out Services Combination, generate the service call sequence; If the Services Combination failure then sends to the abnormality processing submodule and handles;

The customization memory module, function that storage is corresponding and service call sequence are to template base;

Implementation sub-module, according to the master data inquiry service in the sequence call master data inquiry service storehouse of service call sequence, and input/output relation between coordination service, treat that whole service call sequence is finished, the output net result; If unusual, then send to the abnormality processing submodule and handle;

The result generates submodule, and the net result of implementation sub-module output is generated final data query analysis result according to the described display format as a result of request,, carries out the result and shows to customizable inquiring client terminal echo reply with the form of browser page;

The abnormality processing submodule generates failure information, and returns to customizable inquiring client terminal demonstration.

12. according to the described a kind of customizable searching platform based on Services Combination of claim 11, it is characterized in that: in the data object model storehouse, described field, it is according to OO thought that data in the field are carried out abstract classification, and data are carried out feature extraction; The attribute that the field data object model that takes out is comprised is as the attribute of field data object model, but each field data object model is written as the edit control of customizable inquiring client terminal in the mode of client script, describes the data query demand of self to the data query server.

13. according to the described a kind of customizable searching platform of claim 12 based on Services Combination, it is characterized in that: data object model storehouse, described field comprises logical control, with computing and logical operation the incidence relation between the data object attribute of two or more fields is described, each logical control is write with script mode.

14. according to the described a kind of customizable searching platform of claim 11 based on Services Combination, it is characterized in that: the master data inquiry service in described master data inquiry service storehouse is corresponding one by one with the field data object model in data object model storehouse, field, the atomic operation of field data object inquiry is provided, only by one or several attribute of corresponding field data object model, acquisition has the set of the data object example of this attribute from data query data in server storehouse.

15. according to the described a kind of customizable searching platform of claim 14 based on Services Combination, it is characterized in that: each master data inquiry service in described master data inquiry service storehouse comprises the atomic operation of an inquiry relating attribute, inquiry obtains the associated data object and the attribute thereof of corresponding field data object, the community set of output associated data object and associated data object.

16., it is characterized in that: the customization data inquiry that described customizable inquiring client terminal carries out according to the described a kind of customizable searching platform of claim 11 based on Services Combination:

But on the page of customizable inquiring client terminal, from the control in domain data model storehouse, select each edit control, specify each query contents data object of this inquiry; But select the row of the attribute of each edit control as the result data tabulation of this data query;

But on customized web page, from the visual control in domain data model storehouse, select each edit control, specify the condition data object of this inquiry; But property value is selected, filled in to the attribute of each edit control of selection also for this attribute, and the query contents data object of specifying this querying condition of selecting to be used for;

Arrange the sequencing of the control property of selecting, fill customized web page result's data list gauge outfit, and the row that are used as sort key of designated result tabulation;

The data query of the new customization of name is preserved this customization, and names the corresponding template of this customization.

17. according to the described a kind of customizable searching platform of claim 11 based on Services Combination, it is characterized in that: the original state that the described request analyzing sub-module obtains, form by query script set, Query Result and service calling sequence, wherein, the query script set comprises that this user asks the query script of each data object correspondence of needs inquiry, and Query Result and service calling sequence are empty.

18. according to the described a kind of customizable searching platform of claim 11 based on Services Combination, it is characterized in that: the dbjective state that the described request analyzing sub-module obtains, form by query script set, Query Result and service calling sequence, wherein, the query script set is for empty, Query Result is the target data object of each query script in the query script set, and the service call sequence is the memory headroom of one period free time.

19. according to the described a kind of customizable searching platform of claim 11 based on Services Combination, it is characterized in that: described Services Combination engine extracts the source state of this Services Combination process, comprises target data object, querying condition set and the empty service call sequence of query script; Interpolation source state is to the upstate formation that is used for temporary storaging data;

If the upstate formation is empty, then export the Services Combination failure information, the Services Combination process finishes; Otherwise:

From the upstate formation, take out and be positioned at the upstate of queue heads as current state;

If the querying condition of current state is empty, then export the service call sequence of current state, the Services Combination process finishes; Otherwise the querying condition of current state is stored in the interim formation of storing queries condition;

If interim formation be a sky, whether be that sky is judged and handled to the upstate formation; Otherwise:

From interim formation, take out the querying condition conduct that is positioned at queue heads and work as precondition, search can be used for when the master data inquiry service of precondition and search its corresponding atomic operation, whether used if all master data inquiry services are all modulated, be that sky is judged and handled to the upstate formation then; Otherwise:

The atomic operation that finds is added to the service call sequence of current state with relation of equivalence;

At the current querying condition of current state, generate the follow-up state of current state, and follow-up state is added the upstate formation, whether be that sky is judged and handled to interim formation.

20. according to the described a kind of customizable searching platform of claim 19 based on Services Combination, it is characterized in that: the Services Combination process of described Services Combination engine generates the follow-up state of current state, if when the field data object of precondition correspondence is the destination object of this query script, then precondition is worked as in deletion from current state, and the state that will delete behind precondition adds the upstate formation as follow-up state, and generation finishes; Otherwise, call correlation inquiry atomic operation when the master data inquiry service of precondition correspondence, each relating attribute and property value source is added in the set of acquisition relating attribute respectively, and deletion adds the upstate formation when precondition and as follow-up state from current state; Then handle next relating attribute and all follow-up states are added the upstate formation, all relating attributes have all generated follow-up state in the relating attribute set.

Images