From 4e3abe4616aa5c7023f257fe2f947b193cc7f61d Mon Sep 17 00:00:00 2001
From: Thomas Petit-Jean <thomas.petitjean@aphp.fr>
Date: Tue, 21 Jan 2025 11:52:36 +0100
Subject: [PATCH] feat: accept pd.DataFrame in session.createDataFrame

---
 sqlframe/base/session.py                | 28 ++++++++++++++++++-------
 tests/integration/test_int_dataframe.py | 13 ++++++++++++
 2 files changed, 33 insertions(+), 8 deletions(-)

diff --git a/sqlframe/base/session.py b/sqlframe/base/session.py
index f2ddee8..340b632 100644
--- a/sqlframe/base/session.py
+++ b/sqlframe/base/session.py
@@ -2,6 +2,7 @@
 
 from __future__ import annotations
 
+import contextlib
 import datetime
 import logging
 import sys
@@ -202,13 +203,16 @@ def range(self, *args):
 
     def createDataFrame(
         self,
-        data: t.Sequence[
-            t.Union[
-                t.Dict[str, ColumnLiterals],
-                t.List[ColumnLiterals],
-                t.Tuple[ColumnLiterals, ...],
-                ColumnLiterals,
-            ]
+        data: t.Union[
+            t.Sequence[
+                t.Union[
+                    t.Dict[str, ColumnLiterals],
+                    t.List[ColumnLiterals],
+                    t.Tuple[ColumnLiterals, ...],
+                    ColumnLiterals,
+                ],
+            ],
+            pd.DataFrame,
         ],
         schema: t.Optional[SchemaInput] = None,
         samplingRatio: t.Optional[float] = None,
@@ -229,11 +233,18 @@ def createDataFrame(
         ):
             raise NotImplementedError("Only schema of either list or string of list supported")
 
+        with contextlib.suppress(ImportError):
+            from pandas import DataFrame as pd_DataFrame
+
+            if isinstance(data, pd_DataFrame):
+                data = data.to_dict("records")  # type: ignore
+
         column_mapping: t.Mapping[str, t.Optional[exp.DataType]]
         if schema is not None:
             column_mapping = get_column_mapping_from_schema_input(
                 schema, dialect=self.input_dialect
             )
+
         elif data:
             if isinstance(data[0], Row):
                 column_mapping = {col_name.strip(): None for col_name in data[0].__fields__}
@@ -375,7 +386,8 @@ def sql(
         dialect = Dialect.get_or_raise(dialect or self.input_dialect)
         expression = (
             sqlglot.parse_one(
-                normalize_string(sqlQuery, from_dialect=dialect, is_query=True), read=dialect
+                normalize_string(sqlQuery, from_dialect=dialect, is_query=True),
+                read=dialect,
             )
             if isinstance(sqlQuery, str)
             else sqlQuery
diff --git a/tests/integration/test_int_dataframe.py b/tests/integration/test_int_dataframe.py
index 89d0144..5e24e41 100644
--- a/tests/integration/test_int_dataframe.py
+++ b/tests/integration/test_int_dataframe.py
@@ -2,6 +2,7 @@
 
 import typing as t
 
+import pandas as pd
 import pytest
 from _pytest.fixtures import FixtureRequest
 from pyspark.sql import DataFrame as PySparkDataFrame
@@ -29,6 +30,18 @@ def test_empty_df(
     compare_frames(df_empty, dfs_empty, no_empty=False)
 
 
+def test_dataframe_from_pandas(
+    pyspark_employee: PySparkDataFrame,
+    compare_frames: t.Callable,
+):
+    compare_frames(
+        pyspark_employee,
+        pyspark_employee.sparkSession.createDataFrame(
+            pyspark_employee.toPandas(),
+        ),
+    )
+
+
 def test_simple_select(
     pyspark_employee: PySparkDataFrame,
     get_df: t.Callable[[str], BaseDataFrame],