reuse config

python/pyspark/sql/pandas/conversion.py

@@ -785,8 +785,9 @@ def _create_from_arrow_table(
         if not isinstance(schema, StructType):
             schema = from_arrow_schema(table.schema, prefer_timestamp_ntz=prefer_timestamp_ntz)
 
+        safecheck = self._jconf.arrowSafeTypeConversion()
         table = _check_arrow_table_timestamps_localize(table, schema, True, timezone).cast(
-            to_arrow_schema(schema, error_on_duplicated_field_names_in_struct=True)
+            to_arrow_schema(schema, error_on_duplicated_field_names_in_struct=True), safe=safecheck
         )
 
         # Chunk the Arrow Table into RecordBatches

sql/catalyst/src/main/scala/org/apache/spark/sql/internal/SQLConf.scala

@@ -3353,9 +3353,10 @@ object SQLConf {
     buildConf("spark.sql.execution.pandas.convertToArrowArraySafely")
       .internal()
       .doc("When true, Arrow will perform safe type conversion when converting " +
-        "Pandas.Series to Arrow array during serialization. Arrow will raise errors " +
-        "when detecting unsafe type conversion like overflow. When false, disabling Arrow's type " +
-        "check and do type conversions anyway. This config only works for Arrow 0.11.0+.")
+        "Pandas.Series to Arrow array during serialization, and when casting Arrow tables to " +
+        "create DataFrames. Arrow will raise errors when detecting unsafe type conversion like " +
+        "overflow. When false, disabling Arrow's type check and do type conversions anyway. This " +
+        "config only works for Arrow 0.11.0+.")
       .version("3.0.0")
       .booleanConf
       .createWithDefault(false)