[
https://issues.apache.org/jira/browse/IMPALA-13641?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Riza Suminto updated IMPALA-13641:
----------------------------------
Description: While working on IMPALA-13636, I notice that
ParquetUncompressedBytesReadPerColumn and ParquetCompressedBytesReadPerColumn
exist in runtime profile even when no parquet file is read (all scan text
files). These counters should be lazily created only if
HdfsScanNodeBase::bytes_read_per_col_ is not empty. (was: While working on
IMPALA-13636, I notice that ParquetUncompressedBytesReadPerColumn and
ParquetCompressedBytesReadPerColumn exist in runtime profile even when no
parquet file is being (all scan text files). These counters should be lazily
created only if HdfsScanNodeBase::bytes_read_per_col_ is not empty.)
> Parquet specific counters should not exist if not reading any parquet file
> --------------------------------------------------------------------------
>
> Key: IMPALA-13641
> URL: https://issues.apache.org/jira/browse/IMPALA-13641
> Project: IMPALA
> Issue Type: Improvement
> Components: Backend
> Affects Versions: Impala 4.4.0
> Reporter: Riza Suminto
> Assignee: Riza Suminto
> Priority: Minor
>
> While working on IMPALA-13636, I notice that
> ParquetUncompressedBytesReadPerColumn and ParquetCompressedBytesReadPerColumn
> exist in runtime profile even when no parquet file is read (all scan text
> files). These counters should be lazily created only if
> HdfsScanNodeBase::bytes_read_per_col_ is not empty.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
