add Binary String Functions (apache#494)

jiangzhx · web-flow · commit 106786ab57cf · 2023-09-20T05:15:45.000-06:00
diff --git a/datafusion/tests/test_functions.py b/datafusion/tests/test_functions.py
@@ -477,3 +477,17 @@ def test_case(df):
     assert result.column(0) == pa.array([10, 8, 8])
     assert result.column(1) == pa.array(["Hola", "Mundo", "!!"])
     assert result.column(2) == pa.array(["Hola", "Mundo", None])
+
+
+def test_binary_string_functions(df):
+    df = df.select(
+        f.encode(column("a"), literal("base64")),
+        f.decode(f.encode(column("a"), literal("base64")), literal("base64")),
+    )
+    result = df.collect()
+    assert len(result) == 1
+    result = result[0]
+    assert result.column(0) == pa.array(["SGVsbG8", "V29ybGQ", "IQ"])
+    assert pa.array(result.column(1)).cast(pa.string()) == pa.array(
+        ["Hello", "World", "!"]
+    )
diff --git a/src/functions.rs b/src/functions.rs
@@ -332,6 +332,9 @@ scalar_function!(r#struct, Struct); // Use raw identifier since struct is a keyw
 scalar_function!(from_unixtime, FromUnixtime);
 scalar_function!(arrow_typeof, ArrowTypeof);
 scalar_function!(random, Random);
+//Binary String Functions
+scalar_function!(encode, Encode);
+scalar_function!(decode, Decode);
 
 aggregate_function!(approx_distinct, ApproxDistinct);
 aggregate_function!(approx_median, ApproxMedian);
@@ -486,5 +489,9 @@ pub(crate) fn init_module(m: &PyModule) -> PyResult<()> {
     m.add_wrapped(wrap_pyfunction!(var_pop))?;
     m.add_wrapped(wrap_pyfunction!(var_samp))?;
     m.add_wrapped(wrap_pyfunction!(window))?;
+
+    //Binary String Functions
+    m.add_wrapped(wrap_pyfunction!(encode))?;
+    m.add_wrapped(wrap_pyfunction!(decode))?;
     Ok(())
 }
