sqlframe 3.31.3__tar.gz → 3.32.0__tar.gz

{sqlframe-3.31.3 → sqlframe-3.32.0}/.gitignore

@@ -142,4 +142,5 @@ spark_warehouse
 sqlframe/_version.py
 
 # Emacs files
-*~
+*~
+**/.claude/settings.local.json

{sqlframe-3.31.3 → sqlframe-3.32.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: sqlframe
-Version: 3.31.3
+Version: 3.32.0
 Summary: Turning PySpark Into a Universal DataFrame API
 Home-page: https://github.com/eakmanrq/sqlframe
 Author: Ryan Eakman

{sqlframe-3.31.3 → sqlframe-3.32.0}/setup.py

@@ -20,7 +20,7 @@ setup(
     python_requires=">=3.9",
     install_requires=[
         "prettytable<4",
-        "sqlglot>=24.0.0,<26.17",
+        "sqlglot>=24.0.0,<26.18",
         "typing_extensions",
     ],
     extras_require={

{sqlframe-3.31.3 → sqlframe-3.32.0}/sqlframe/_version.py

@@ -17,5 +17,5 @@ __version__: str
 __version_tuple__: VERSION_TUPLE
 version_tuple: VERSION_TUPLE
 
-__version__ = version = '3.31.3'
-__version_tuple__ = version_tuple = (3, 31, 3)
+__version__ = version = '3.32.0'
+__version_tuple__ = version_tuple = (3, 32, 0)

{sqlframe-3.31.3 → sqlframe-3.32.0}/sqlframe/base/dataframe.py

@@ -260,6 +260,13 @@ class BaseDataFrame(t.Generic[SESSION, WRITER, NA, STAT, GROUP_DATA]):
     def __copy__(self):
         return self.copy()
 
+    def __repr__(self) -> str:
+        fields = [f"{field.name}: {field.dataType}" for field in self.schema]
+        return "DataFrame[" + ", ".join(fields) + "]"
+
+    def _display_(self) -> str:
+        return self.__repr__()
+
     @property
     def _typed_columns(self) -> t.List[CatalogColumn]:
         raise NotImplementedError

{sqlframe-3.31.3 → sqlframe-3.32.0}/sqlframe/base/functions.py

@@ -494,21 +494,22 @@ def skewness(col: ColumnOrName) -> Column:
         func_name = "SKEW"
 
     if session._is_duckdb or session._is_snowflake:
+        col = Column.ensure_col(col)
         when_func = get_func_from_session("when")
         count_func = get_func_from_session("count")
-        count_star = count_func("*")
+        count_col = count_func(col)
         lit_func = get_func_from_session("lit")
         sqrt_func = get_func_from_session("sqrt")
-        col = Column.ensure_col(col)
+        full_calc = (
+            Column.invoke_anonymous_function(col, func_name)
+            * (count_col - lit_func(2))
+            / (sqrt_func(count_col * (count_col - lit_func(1))))
+        )
         return (
-            when_func(count_star == lit_func(0), lit_func(None))
-            .when(count_star == lit_func(1), lit_func(float("nan")))
-            .when(count_star == lit_func(2), lit_func(0.0))
-            .otherwise(
-                Column.invoke_anonymous_function(col, func_name)
-                * (count_star - lit_func(2))
-                / (sqrt_func(count_star * (count_star - lit_func(1))))
-            )
+            when_func(count_col == lit_func(0), lit_func(None))
+            .when(count_col == lit_func(1), lit_func(None))
+            .when(count_col == lit_func(2), lit_func(0.0))
+            .otherwise(full_calc)
         )
 
     return Column.invoke_anonymous_function(col, func_name)

{sqlframe-3.31.3 → sqlframe-3.32.0}/sqlframe/base/session.py

@@ -304,8 +304,6 @@ class _BaseSession(t.Generic[CATALOG, READER, WRITER, DF, TABLE, CONN, UDF_REGIS
             elif isinstance(value, float):
                 return "double"
             elif isinstance(value, datetime.datetime):
-                if value.tzinfo:
-                    return "timestamptz"
                 return "timestamp"
             elif isinstance(value, datetime.date):
                 return "date"

{sqlframe-3.31.3 → sqlframe-3.32.0}/sqlframe/base/util.py

@@ -313,7 +313,7 @@ def sqlglot_to_spark(sqlglot_dtype: exp.DataType) -> types.DataType:
         exp.DataType.Type.TIMESTAMP: types.TimestampType,
         exp.DataType.Type.TIMESTAMPTZ: types.TimestampType,
         exp.DataType.Type.TIMESTAMPLTZ: types.TimestampType,
-        exp.DataType.Type.TIMESTAMPNTZ: types.TimestampType,
+        exp.DataType.Type.TIMESTAMPNTZ: types.TimestampNTZType,
         exp.DataType.Type.DATE: types.DateType,
         exp.DataType.Type.JSON: types.StringType,
     }

{sqlframe-3.31.3 → sqlframe-3.32.0}/sqlframe.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: sqlframe
-Version: 3.31.3
+Version: 3.32.0
 Summary: Turning PySpark Into a Universal DataFrame API
 Home-page: https://github.com/eakmanrq/sqlframe
 Author: Ryan Eakman

{sqlframe-3.31.3 → sqlframe-3.32.0}/sqlframe.egg-info/requires.txt

@@ -1,5 +1,5 @@
 prettytable<4
-sqlglot<26.17,>=24.0.0
+sqlglot<26.18,>=24.0.0
 typing_extensions
 
 [bigquery]

{sqlframe-3.31.3 → sqlframe-3.32.0}/tests/integration/engines/test_int_functions.py

@@ -692,6 +692,10 @@ def test_skewness(get_session_and_func):
     session, skewness = get_session_and_func("skewness")
     df = session.createDataFrame([[1], [1], [2]], ["c"])
     assert math.isclose(df.select(skewness(df.c)).first()[0], 0.7071067811865475, rel_tol=1e-5)
+    df = session.createDataFrame([{"a": 2.0}, {"a": None}])
+    assert df.agg(skewness(df.a)).collect() == [Row(value=None)]
+    df = session.createDataFrame([{"a": 1}, {"a": 2}])
+    assert df.agg(skewness(df.a)).collect() == [Row(value=0.0)]
 
 
 def test_kurtosis(get_session_and_func):
@@ -1423,6 +1427,11 @@ def test_to_timestamp(get_session_and_func):
     assert result == datetime.datetime(1997, 2, 28, 10, 30)
     result = df.select(to_timestamp(df.t, "yyyy-MM-dd HH:mm:ss").alias("dt")).first()[0]
     assert result == datetime.datetime(1997, 2, 28, 10, 30)
+    df = session.createDataFrame(pd.DataFrame({"a": ["2020-01-01 01:02:03+0100"]}))
+    # There seems to be an issue in SQLGlot transpiling the format string for these dialects
+    if not isinstance(session, (BigQuerySession, SnowflakeSession, PostgresSession)):
+        result = df.withColumn("b", to_timestamp("a", format="y-M-d H:m:sZ")).first()[0]
+        assert result == "2020-01-01 01:02:03+0100"
 
 
 def test_trunc(get_session_and_func):