n3nash commented on a change in pull request #674: Upgrade to Hive 2.x, MOR
read query fixes and performance improvement
URL: https://github.com/apache/incubator-hudi/pull/674#discussion_r292099587
##########
File path:
hoodie-hadoop-mr/src/main/java/com/uber/hoodie/hadoop/realtime/AbstractRealtimeRecordReader.java
##########
@@ -262,8 +280,8 @@ public static Writable avroToArrayWritable(Object value,
Schema schema) {
mapValues[1] = avroToArrayWritable(mapEntry.getValue(),
schema.getValueType());
values3[index3++] = new ArrayWritable(Writable.class, mapValues);
}
- wrapperWritable = new Writable[]{new ArrayWritable(Writable.class,
values3)};
- return new ArrayWritable(Writable.class, wrapperWritable);
+ // Hive 1.x will fail here, it requires values3 to be wrapped into
another ArrayWritable
Review comment:
We could, I'm not sure if Hive 2.x is backwards compatible with Hive 1.x for
on-the-wire stuff. Going forward we are dropping support for Hive 1.x which
means we give no guarantee that clients can compile against hive 1.x. So, a
Hive 2.x client has to be able to work with Hive 1.x server
(metastore/hiveserver) - I haven't tested this case since doesn't seem worth
it. Let me know.
----------------------------------------------------------------
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:
[email protected]
With regards,
Apache Git Services
