Github user SongYadong commented on a diff in the pull request: https://github.com/apache/spark/pull/22348#discussion_r215870108 --- Diff: sql/core/src/main/java/org/apache/spark/sql/execution/datasources/parquet/VectorizedParquetRecordReader.java --- @@ -154,8 +159,6 @@ public void close() throws IOException { @Override public boolean nextKeyValue() throws IOException { - resultBatch(); --- End diff -- Thanks for your review. calling sequence is optional. `initBatch()` was moved to `nextBatch()`, if `hasNext()` called before `initBatch()`, eventually `initBatch()` will be called first of all in `nextBatch()`: ``` public boolean nextBatch() throws IOException { if (columnarBatch == null) initBatch(); ``` Orc has a much simple way: ``` public boolean nextKeyValue() throws IOException { return nextBatch(); } ```
--- --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org