[
https://issues.apache.org/jira/browse/PARQUET-1381?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17043505#comment-17043505
]
Gabor Szadovszky commented on PARQUET-1381:
-------------------------------------------
I don't think anyone is working on it. Feel free to open a PR and I'm happy to
review.
However, when I had to revert the previous implementation I was thinking a
correct solution. It turned out that it's so complicated to be optimal in
performance yet correct from statistics point of view. Currently, I am not sure
if it worth the potential efforts to implement this feature while you can get a
similar functionality by rewriting a whole table of parquet files in a
distributed way using a query engine (e.g. Hive/Impala/Spark). Also, you can do
it as a background process using the Hive compaction feature.
> Add merge blocks command to parquet-tools
> -----------------------------------------
>
> Key: PARQUET-1381
> URL: https://issues.apache.org/jira/browse/PARQUET-1381
> Project: Parquet
> Issue Type: New Feature
> Components: parquet-mr
> Affects Versions: 1.10.0
> Reporter: Ekaterina Galieva
> Assignee: Ekaterina Galieva
> Priority: Major
> Labels: pull-request-available
>
> Current implementation of merge command in parquet-tools doesn't merge row
> groups, just places one after the other. Add API and command option to be
> able to merge small blocks into larger ones up to specified size limit.
> h6. Implementation details:
> Blocks are not reordered not to break possible initial predicate pushdown
> optimizations.
> Blocks are not divided to fit upper bound perfectly.
> This is an intentional performance optimization.
> This gives an opportunity to form new blocks by coping full content of
> smaller blocks by column, not by row.
> h6. Examples:
> # Input files with blocks sizes:
> {code:java}
> [128 | 35], [128 | 40], [120]{code}
> Expected output file blocks sizes:
> {{merge }}
> {code:java}
> [128 | 35 | 128 | 40 | 120]
> {code}
> {{merge -b}}
> {code:java}
> [128 | 35 | 128 | 40 | 120]
> {code}
> {{merge -b -l 256 }}
> {code:java}
> [163 | 168 | 120]
> {code}
> # Input files with blocks sizes:
> {code:java}
> [128 | 35], [40], [120], [6] {code}
> Expected output file blocks sizes:
> {{merge}}
> {code:java}
> [128 | 35 | 40 | 120 | 6]
> {code}
> {{merge -b}}
> {code:java}
> [128 | 75 | 126]
> {code}
> {{merge -b -l 256}}
> {code:java}
> [203 | 126]{code}
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
