marton-bod commented on a change in pull request #3131:
URL: https://github.com/apache/hive/pull/3131#discussion_r840411036
##########
File path:
iceberg/iceberg-handler/src/main/java/org/apache/iceberg/mr/hive/HiveIcebergOutputFormat.java
##########
@@ -83,9 +83,19 @@ private static HiveIcebergRecordWriter writer(JobConf jc) {
.operationId(operationId)
.build();
String tableName = jc.get(Catalogs.NAME);
- HiveFileWriterFactory hfwf = new HiveFileWriterFactory(table, fileFormat,
schema,
- null, fileFormat, null, null, null, null);
- return new HiveIcebergRecordWriter(schema, spec, fileFormat,
- hfwf, outputFileFactory, io, targetFileSize, taskAttemptID, tableName);
+ if (HiveIcebergStorageHandler.isDelete(jc, tableName)) {
+ // TODO: remove this Avro-specific logic once we have Avro writer
function ready
+ // for now, this means that Avro delete files will not contain the 'row'
column
+ Schema positionDeleteRowSchema = fileFormat == FileFormat.AVRO ? null :
schema;
+ HiveFileWriterFactory hfwf = new HiveFileWriterFactory(table,
fileFormat, schema,
+ null, fileFormat, null, null, null, positionDeleteRowSchema);
+ return new HiveIcebergDeleteWriter(hfwf, schema, spec, fileFormat,
outputFileFactory, io, targetFileSize,
+ taskAttemptID, tableName);
Review comment:
hmm, I did it this way mostly because the hfwf in the delete case needs
the additional `positionDeleteRowSchema ` populated. However, I guess we could
supply the delete row schema for the simple write path too, it just wouldn't be
used. I'll give it a try
--
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: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
