sqlframe 3.38.1__tar.gz → 3.38.2__tar.gz

{sqlframe-3.38.1 → sqlframe-3.38.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: sqlframe
-Version: 3.38.1
+Version: 3.38.2
 Summary: Turning PySpark Into a Universal DataFrame API
 Home-page: https://github.com/eakmanrq/sqlframe
 Author: Ryan Eakman

{sqlframe-3.38.1 → sqlframe-3.38.2}/sqlframe/_version.py

@@ -17,5 +17,5 @@ __version__: str
 __version_tuple__: VERSION_TUPLE
 version_tuple: VERSION_TUPLE
 
-__version__ = version = '3.38.1'
-__version_tuple__ = version_tuple = (3, 38, 1)
+__version__ = version = '3.38.2'
+__version_tuple__ = version_tuple = (3, 38, 2)

{sqlframe-3.38.1 → sqlframe-3.38.2}/sqlframe/base/functions.py

@@ -2318,6 +2318,14 @@ def array_distinct(col: ColumnOrName) -> Column:
     if session._is_bigquery:
         return array_distinct_bgutil(col)
 
+    if session._is_duckdb:
+        # DuckDB's array_distinct removes nulls, but we need to preserve them
+        # Check if original array contains null and append it back if needed
+        original_col = Column.ensure_col(col)
+        distinct_result = Column.invoke_anonymous_function(col, "ARRAY_DISTINCT")
+        has_null = array_position(original_col, lit(None)) > lit(0)
+        return when(has_null, array_append(distinct_result, lit(None))).otherwise(distinct_result)
+
     return Column.invoke_anonymous_function(col, "ARRAY_DISTINCT")
 
 

{sqlframe-3.38.1 → sqlframe-3.38.2}/sqlframe.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: sqlframe
-Version: 3.38.1
+Version: 3.38.2
 Summary: Turning PySpark Into a Universal DataFrame API
 Home-page: https://github.com/eakmanrq/sqlframe
 Author: Ryan Eakman

{sqlframe-3.38.1 → sqlframe-3.38.2}/tests/integration/engines/test_int_functions.py

@@ -2290,16 +2290,21 @@ def test_array_remove(get_session_and_func, get_func):
 def test_array_distinct(get_session_and_func, get_func):
     session, array_distinct = get_session_and_func("array_distinct")
     lit = get_func("lit", session)
+    # BigQuery doesn't support null values in arrays
+    value3 = lit([1]) if isinstance(session, BigQuerySession) else lit([1, None])
     results = (
         session.range(1)
         .select(
             array_distinct(lit([1, 2, 3, 2])).alias("value"),
             array_distinct(lit([4, 5, 5, 4])).alias("value2"),
+            array_distinct(lit(value3)).alias("value3"),
         )
         .collect()
     )
     assert results[0][0] in ([1, 2, 3], [3, 2, 1])
     assert results[0][1] in ([4, 5], [5, 4])
+    if not isinstance(session, BigQuerySession):
+        assert results[0][2] in ([1, None], [None, 1])
 
 
 def test_array_intersect(get_session_and_func, get_func):