[
https://issues.apache.org/jira/browse/HIVE-21218?focusedWorklogId=196111&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-196111
]
ASF GitHub Bot logged work on HIVE-21218:
-----------------------------------------
Author: ASF GitHub Bot
Created on: 08/Feb/19 08:02
Start Date: 08/Feb/19 08:02
Worklog Time Spent: 10m
Work Description: xbaran commented on pull request #526: HIVE-21218:
KafkaSerDe doesn't support topics created via Confluent
URL: https://github.com/apache/hive/pull/526#discussion_r254981251
##########
File path: kafka-handler/src/java/org/apache/hadoop/hive/kafka/KafkaSerDe.java
##########
@@ -369,6 +379,20 @@ private SubStructObjectInspector(StructObjectInspector
baseOI, int toIndex) {
}
}
+ static class ConfluentAvroBytesConverter extends AvroBytesConverter {
+ ConfluentAvroBytesConverter(Schema schema) {
+ super(schema);
+ }
+
+ @Override
+ Decoder getDecoder(byte[] value) {
+ /**
+ * Confluent 4 magic bytes that represents Schema ID as Integer. These
bits are added before value bytes.
+ */
+ return DecoderFactory.get().binaryDecoder(value, 5, value.length - 5,
null);
Review comment:
The problem is you have to provide schema during `initialize` and there is
no record to read schema ID from at that time. I was thinking about switch
`bytesConverter` to a function/supplier or factory and give it to the
KafkaWritable and resolve byteconverter lazy way. However I'm quite new to Hive
and need to deep more to mechanics how the low level api works. I'd like to see
hive download schemas from schema registry, but I need to think about concept
first.
About the compatibility, there are several compatibility modes in schema
registry. It can be set for each topic and you can set BACKWARD compat for
example.
And yea you can have objects with literally different schemas not just
different versions in one topic and I think this problem Hive cannot address.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on 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
Issue Time Tracking
-------------------
Worklog Id: (was: 196111)
Time Spent: 10m
Remaining Estimate: 0h
> KafkaSerDe doesn't support topics created via Confluent Avro serializer
> -----------------------------------------------------------------------
>
> Key: HIVE-21218
> URL: https://issues.apache.org/jira/browse/HIVE-21218
> Project: Hive
> Issue Type: Bug
> Components: kafka integration, Serializers/Deserializers
> Affects Versions: 3.1.1
> Reporter: Milan Baran
> Assignee: Milan Baran
> Priority: Major
> Labels: pull-request-available
> Time Spent: 10m
> Remaining Estimate: 0h
>
> According to [Google
> groups|https://groups.google.com/forum/#!topic/confluent-platform/JYhlXN0u9_A]
> the Confluent avro serialzier uses propertiary format for kafka value -
> <magic_byte 0x00><4 bytes of schema ID><regular avro bytes for object that
> conforms to schema>.
> This format does not cause any problem for Confluent kafka deserializer which
> respect the format however for hive kafka handler its bit a problem to
> correctly deserialize kafka value, because Hive uses custom deserializer from
> bytes to objects and ignores kafka consumer ser/deser classes provided via
> table property.
> It would be nice to support Confluent format with magic byte.
> Also it would be great to support Schema registry as well.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
