[ https://issues.apache.org/jira/browse/KAFKA-6623?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Jason Gustafson updated KAFKA-6623: ----------------------------------- Description: Certain batch-level operations are only efficient with the new message format version. For example, {{RecordBatch.baseOffset}} requires decompression for the old message format versions. It is a bit too easy at the moment to overlook the performance implications when using these methods, which results in issues like KAFKA-6622. We should consider either renaming them to make the complexity more apparent or modify the API so that the old message format simply expose the option conveniently. A similar case is the record count, which is efficient in v2, but inefficient for older formats. We handled this case by exposing a {{countOrNull}} method which only returns the count for v2. (was: Certain batch-level operations are only efficient with the new message format version. For example, {{RecordBatch.baseOffset}} requires decompression for the old message format versions. It is a bit too easy at the moment to overlook the performance implications when using these methods at the moment, which results in issues like KAFKA-6622. We should consider either renaming them to make the complexity more apparent or modify the API so that the old message format simply expose the option conveniently. A similar case is the record count, which is efficient in v2, but inefficient for older formats. We handled this case by exposing a {{countOrNull}} method which only returns the count for v2.) > Consider renaming inefficient RecordBatch operations > ---------------------------------------------------- > > Key: KAFKA-6623 > URL: https://issues.apache.org/jira/browse/KAFKA-6623 > Project: Kafka > Issue Type: Improvement > Reporter: Jason Gustafson > Priority: Major > > Certain batch-level operations are only efficient with the new message format > version. For example, {{RecordBatch.baseOffset}} requires decompression for > the old message format versions. It is a bit too easy at the moment to > overlook the performance implications when using these methods, which results > in issues like KAFKA-6622. We should consider either renaming them to make > the complexity more apparent or modify the API so that the old message format > simply expose the option conveniently. A similar case is the record count, > which is efficient in v2, but inefficient for older formats. We handled this > case by exposing a {{countOrNull}} method which only returns the count for v2. -- This message was sent by Atlassian JIRA (v7.6.3#76005)