Re: [PR] [SPARK-45908][Python] Add support for writing empty DataFrames to parquet with partitions [spark]

["HyukjinKwon (via GitHub)" <gi...@apache.org>]

HyukjinKwon commented on code in PR #43798:
URL: https://github.com/apache/spark/pull/43798#discussion_r1393546849


##########
python/pyspark/sql/readwriter.py:
##########
@@ -1936,7 +1936,23 @@ def parquet(
         if partitionBy is not None:
             self.partitionBy(partitionBy)
         self._set_opts(compression=compression)
-        self._jwrite.parquet(path)
+        if self._df.isEmpty() is True and partitionBy is not None:
+            self._write_empty_partition(path, partitionBy)
+        else:
+            self._jwrite.parquet(path)
+    
+    @overload
+    def _write_empty_partition(self, path: str, partitionBy: str) -> None:
+        ...
+    
+    @overload
+    def _write_empty_partition(self, path: str, partitionBy: List[str]) -> None:
+        ...
+
+    def _write_empty_partition(self, path: str, partitionBy: Union[str, List[str]]) -> None:
+            temp_row_data = {col: None for col in partitionBy} if isinstance(partitionBy, list) else {partitionBy: None}
+            temp_df = self._df.sparkSession.createDataFrame([Row(**temp_row_data)])
+            temp_df.write.mode("overwrite").partitionBy(partitionBy).parquet(path)

Review Comment:
   Does this work with Scala API too?



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org