[GitHub] [spark] parthchandra commented on pull request #35262: [SPARK-37974][SQL] Implement vectorized DELTA_BYTE_ARRAY and DELTA_LENGTH_BYTE_ARRAY encodings for Parquet V2 support

[GitBox <gi...@apache.org>]

parthchandra commented on pull request #35262:
URL: https://github.com/apache/spark/pull/35262#issuecomment-1029288528


   @LuciferYang you had a comment that seems to have vanished. From the email - 
   
   > In [sql/core/src/main/java/org/apache/spark/sql/execution/datasources/parquet/VectorizedDeltaByteArrayReader.java](https://github.com/apache/spark/pull/35262#discussion_r797417782):
   > 
   > >        }
   > +      outputWriter.write(c, rowId + i, ByteBuffer.wrap(previous), previous.length);
   > Is line106 always equivalent to c.putByteArray(rowId, previous, 0, length)?
   
   There are multiple output writers. The method called in `readBinary` is equivalent to `c.putByteArray(rowId, previous, 0, length)`; however, if (and when) we add back the OFF_HEAP option, this method will change depending on the memory mode. 
   Also `skipBytes` also calls the same `readValues` method but with a different output writer which does nothing. Note that we still have to process each value even if we do not write it out (because we need the previous value to compute every subsequent value).


-- 
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