sathyaprakashg commented on a change in pull request #2012:
URL: https://github.com/apache/hudi/pull/2012#discussion_r485249027
##########
File path: hudi-common/src/main/java/org/apache/hudi/avro/HoodieAvroUtils.java
##########
@@ -127,12 +128,59 @@ public static GenericRecord bytesToAvro(byte[] bytes,
Schema schema) throws IOEx
* Convert serialized bytes back into avro record.
*/
public static GenericRecord bytesToAvro(byte[] bytes, Schema writerSchema,
Schema readerSchema) throws IOException {
Review comment:
@n3nash Yes, it is very valid point. Adding namespace to reader schema
might make it incompatible with data already written without namespace.
I am thinking we should probably remove the namespace in the writerSchema
fixed fields instead of adding namespace to the reader schema fixed fields.
Since we are not altering reader schema in this approach, there shouldn't be
incompatibility issue with data already written.
If you also agree with this approach, i will update
`AvroConversionHelper.createConverterToAvro` method to remove namespace in
writer schema fixed fields and revoke the changes made in `HoodieAvroUtils`
----------------------------------------------------------------
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]
