[
https://issues.apache.org/jira/browse/HADOOP-1823?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12549129
]
Benjamin Reed commented on HADOOP-1823:
---------------------------------------
We can't really wrap bzip because it is constructed with a byte oriented input
stream. Bzip however processes a bit oriented stream. (Much to my dismay!)
Nothing is byte aligned. Thus the signature, block headers, and the compressed
data do not start on byte boundaries.
There is a non-trivial price to pay to read a bit at a time. In order to wrap
we would have to read the input as bits to find the signature, align the bits
to bytes (shift into new bytes), output a new byte stream, bzip would then
convert those bytes back to bits and process the bit stream. I think bzip has
enough overhead as it is.
I agree that it would be great to submit a patch to Ant to add this
functionality to bzip.
> want InputFormat for bzip2 files
> --------------------------------
>
> Key: HADOOP-1823
> URL: https://issues.apache.org/jira/browse/HADOOP-1823
> Project: Hadoop
> Issue Type: New Feature
> Components: mapred
> Reporter: Doug Cutting
> Attachments: bzip2.jar
>
>
> Unlike gzip, the bzip file format supports splitting. Compression is by
> blocks (900k by default) and blocks are separated by a synchronization marker
> (a 48-bit approximation of Pi). This would permit very large compressed
> files to be split into multiple map tasks, which is not currently possible
> unless using a Hadoop-specific file format.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
