hadoop-common-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jim Kellerman (JIRA)" <j...@apache.org>
Subject [jira] Commented: (HADOOP-3315) New binary file format
Date Tue, 23 Sep 2008 20:07:44 GMT

    [ https://issues.apache.org/jira/browse/HADOOP-3315?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12633869#action_12633869
] 

Jim Kellerman commented on HADOOP-3315:
---------------------------------------

In Data Model/Operations (page 2) it says: "A TFile is not accessible for read while it is
being created." While I can understand why that is reasonable for a sorted TFile since the
index is written last, for a TFile that is unsorted (so it behaves as a SequenceFile) it should
be possible to open a Reader on the TFile to read up to the last sync, provided you do a getFileStatus
to get the "current length". Is this not the case?

If TFile does not support sync, then deprecating SequenceFile (as has been discussed above)
would be a problem for HBase.

I'd also like to see a comparison between TFile's random access performance vs MapFile's random
access performance.


> New binary file format
> ----------------------
>
>                 Key: HADOOP-3315
>                 URL: https://issues.apache.org/jira/browse/HADOOP-3315
>             Project: Hadoop Core
>          Issue Type: New Feature
>          Components: io
>            Reporter: Owen O'Malley
>            Assignee: Amir Youssefi
>         Attachments: HADOOP-3315_20080908_TFILE_PREVIEW_WITH_LZO_TESTS.patch, HADOOP-3315_20080915_TFILE.patch,
TFile Specification Final.pdf
>
>
> SequenceFile's block compression format is too complex and requires 4 codecs to compress
or decompress. It would be good to have a file format that only needs 

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


Mime
View raw message