You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@hudi.apache.org by "sivabalan narayanan (Jira)" <ji...@apache.org> on 2020/04/21 02:06:00 UTC
[jira] [Commented] (HUDI-351) Implement Range + Bloom Filter
checking in one go to improve speed of index
[ https://issues.apache.org/jira/browse/HUDI-351?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17088208#comment-17088208 ]
sivabalan narayanan commented on HUDI-351:
------------------------------------------
[~vinoth]: Do you think we still need to work on this given that we have BloomIndexV2?
> Implement Range + Bloom Filter checking in one go to improve speed of index
> ---------------------------------------------------------------------------
>
> Key: HUDI-351
> URL: https://issues.apache.org/jira/browse/HUDI-351
> Project: Apache Hudi (incubating)
> Issue Type: New Feature
> Components: Index, Performance
> Reporter: Vinoth Chandar
> Priority: Major
>
> Currently, we read the min/max ranges once for range pruning and again read the footer metadata to check for bloom filter..
> Once spark 2.4 support and the 2GB limitations are gone.. worth revisiting if we could do this in a single pass for cases where the bloom filters could fit into memory or implement this check as a RDD operation..
--
This message was sent by Atlassian Jira
(v8.3.4#803005)