liyafan82 commented on a change in pull request #7326:
URL: https://github.com/apache/arrow/pull/7326#discussion_r465618946



##########
File path: 
java/vector/src/main/java/org/apache/arrow/vector/ipc/message/ArrowRecordBatch.java
##########
@@ -232,6 +252,12 @@ public long computeBodyLength() {
       // round up size to the next multiple of 8
       size = DataSizeRoundingUtil.roundUpTo8Multiple(size);
     }
+
+    if (bodyCompression != null) {

Review comment:
       Sounds reasonable. @rymurr had a similar suggestion.
   
   To make the bodyCompression non-null, we need a default compression codec 
that makes no compression/decompression. However, our specification does not 
support such an option now (Please see 
https://github.com/apache/arrow/blob/master/format/Message.fbs#L45-L53). 
Providing one would make the implementation not aligning with the specification.
   
   However, we have revised the PR to provide one, because it seems the C++ 
implementation is not aligning with the specification either (pls see 
https://github.com/apache/arrow/blob/master/cpp/src/arrow/util/compression.h#L33)
   
   After that, we can guarantee that the bodyCompression is never null, and 
some other problems also go away. 
   
   




----------------------------------------------------------------
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.

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


Reply via email to