sqlframe 3.9.0__tar.gz → 3.9.1__tar.gz

{sqlframe-3.9.0 → sqlframe-3.9.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: sqlframe
-Version: 3.9.0
+Version: 3.9.1
 Summary: Turning PySpark Into a Universal DataFrame API
 Home-page: https://github.com/eakmanrq/sqlframe
 Author: Ryan Eakman

{sqlframe-3.9.0 → sqlframe-3.9.1}/sqlframe/_version.py

@@ -12,5 +12,5 @@ __version__: str
 __version_tuple__: VERSION_TUPLE
 version_tuple: VERSION_TUPLE
 
-__version__ = version = '3.9.0'
-__version_tuple__ = version_tuple = (3, 9, 0)
+__version__ = version = '3.9.1'
+__version_tuple__ = version_tuple = (3, 9, 1)

{sqlframe-3.9.0 → sqlframe-3.9.1}/sqlframe/base/dataframe.py

@@ -866,6 +866,7 @@ class _BaseDataFrame(t.Generic[SESSION, WRITER, NA, STAT, GROUP_DATA]):
         if on is None:
             logger.warning("Got no value for on. This appears to change the join to a cross join.")
             how = "cross"
+
         other_df = other_df._convert_leaf_to_cte()
         join_expression = self._add_ctes_to_expression(self.expression, other_df.expression.ctes)
         # We will determine actual "join on" expression later so we don't provide it at first
@@ -875,6 +876,13 @@ class _BaseDataFrame(t.Generic[SESSION, WRITER, NA, STAT, GROUP_DATA]):
         self_columns = self._get_outer_select_columns(join_expression)
         other_columns = self._get_outer_select_columns(other_df.expression)
         join_columns = self._ensure_and_normalize_cols(on)
+        # If the two dataframes being joined come from the same branch and reference the same table,
+        # then really one of the tables was supposed to reference the other dataframe so we update it to do that.
+        if self.branch_id == other_df.branch_id:
+            for col in join_columns:
+                for eq in col.expression.find_all(exp.EQ):
+                    if eq.this.table == eq.expression.table:
+                        eq.expression.set("table", exp.to_identifier(other_df.latest_cte_name))
         # Determines the join clause and select columns to be used passed on what type of columns were provided for
         # the join. The columns returned changes based on how the on expression is provided.
         if how != "cross":

{sqlframe-3.9.0 → sqlframe-3.9.1}/sqlframe.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: sqlframe
-Version: 3.9.0
+Version: 3.9.1
 Summary: Turning PySpark Into a Universal DataFrame API
 Home-page: https://github.com/eakmanrq/sqlframe
 Author: Ryan Eakman

{sqlframe-3.9.0 → sqlframe-3.9.1}/tests/integration/test_int_dataframe.py

@@ -2247,3 +2247,33 @@ def test_chaining_joins_with_selects(
     )
 
     compare_frames(df, dfs, compare_schema=False)
+
+
+# https://github.com/eakmanrq/sqlframe/issues/210
+def test_self_join(
+    pyspark_employee: PySparkDataFrame,
+    get_df: t.Callable[[str], _BaseDataFrame],
+    compare_frames: t.Callable,
+    is_spark: t.Callable,
+):
+    if is_spark():
+        pytest.skip(
+            "This test is not supported in Spark. This is related to how duplicate columns are handled in Spark"
+        )
+    df_filtered = pyspark_employee.where(F.col("age") > 40)
+    df_joined = pyspark_employee.join(
+        df_filtered,
+        pyspark_employee["employee_id"] == df_filtered["employee_id"],
+        how="inner",
+    )
+
+    employee = get_df("employee")
+
+    dfs_filtered = employee.where(SF.col("age") > 40)
+    dfs_joined = employee.join(
+        dfs_filtered,
+        employee["employee_id"] == dfs_filtered["employee_id"],
+        how="inner",
+    )
+
+    compare_frames(df_joined, dfs_joined, compare_schema=False)