crunch-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Gabriel Reid (JIRA)" <j...@apache.org>
Subject [jira] [Updated] (CRUNCH-608) HFile creation doesn't write Bloom filters
Date Tue, 10 May 2016 09:05:12 GMT

     [ https://issues.apache.org/jira/browse/CRUNCH-608?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Gabriel Reid updated CRUNCH-608:
--------------------------------
    Attachment: CRUNCH-608.patch

Patch to use a correctly-configured StoreFile.Writer (instead of HFile.Writer) for writing
HFiles. The StoreFile.Writer includes the necessary logic for writing Bloom filters within
HFiles.

Also includes an addition to existing integration tests to verify that the bloom filter data
is written.

> HFile creation doesn't write Bloom filters
> ------------------------------------------
>
>                 Key: CRUNCH-608
>                 URL: https://issues.apache.org/jira/browse/CRUNCH-608
>             Project: Crunch
>          Issue Type: Bug
>            Reporter: Gabriel Reid
>            Assignee: Gabriel Reid
>         Attachments: CRUNCH-608.patch
>
>
> The HFile writing logic in HFileOutputFormatForCrunch doesn't write Bloom filter data
to the created HFiles, even if bloom filters are enabled on the column family that is being
written (which is the default).
> This means that the created HFiles may offer degraded performance (until they are replaced
via compaction) when searching for rows which don't exist in the given HFiles.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message