Add read API to convert result to Polars

Author: reswqa

dev/dev-requirements.txt

@@ -28,3 +28,4 @@ pytz>=2018.3
 pytest~=7.0
 duckdb>=0.5.0,<2.0.0
 ray~=2.10.0
+polars~=1.15.0

paimon_python_api/table_read.py

@@ -17,6 +17,7 @@
 #################################################################################
 
 import pandas as pd
+import polars as pl
 import pyarrow as pa
 import ray
 
@@ -41,6 +42,10 @@ def to_arrow_batch_reader(self, splits: List[Split]) -> pa.RecordBatchReader:
     def to_pandas(self, splits: List[Split]) -> pd.DataFrame:
         """Read data from splits and converted to pandas.DataFrame format."""
 
+    @abstractmethod
+    def to_polars(self, splits: List[Split]) -> pl.DataFrame:
+        """Read data from splits and converted to polars.DataFrame format."""
+
     @abstractmethod
     def to_duckdb(
             self,

paimon_python_java/pypaimon.py

@@ -18,6 +18,7 @@
 
 import duckdb
 import pandas as pd
+import polars as pl
 import pyarrow as pa
 import ray
 
@@ -164,6 +165,9 @@ def to_arrow_batch_reader(self, splits):
     def to_pandas(self, splits: List[Split]) -> pd.DataFrame:
         return self.to_arrow(splits).to_pandas()
 
+    def to_polars(self, splits: List[Split]) -> pl.DataFrame:
+        return pl.from_arrow(self.to_arrow(splits))
+
     def to_duckdb(
             self,
             splits: List[Split],

paimon_python_java/tests/test_write_and_read.py

@@ -22,8 +22,10 @@
 import unittest
 import pandas as pd
 import pyarrow as pa
+import polars as pl
 from py4j.protocol import Py4JJavaError
 
+from polars import testing as pl_testing
 from paimon_python_api import Schema
 from paimon_python_java import Catalog
 from paimon_python_java.java_gateway import get_gateway
@@ -297,6 +299,10 @@ def testAllWriteAndReadApi(self):
         pd.testing.assert_frame_equal(
             actual.reset_index(drop=True), all_data.reset_index(drop=True))
 
+        # to_polars
+        pl_df = table_read.to_polars(splits)
+        pl_testing.assert_frame_equal(pl_df, pl.from_pandas(all_data))
+
         # to_duckdb
         duckdb_con = table_read.to_duckdb(splits, 'duckdb_table')
         # select *