hadoop-common-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Hairong Kuang (JIRA)" <j...@apache.org>
Subject [jira] Commented: (HADOOP-3498) File globbing alternation should be able to span path components
Date Wed, 13 Aug 2008 17:58:46 GMT

    [ https://issues.apache.org/jira/browse/HADOOP-3498?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12622297#action_12622297
] 

Hairong Kuang commented on HADOOP-3498:
---------------------------------------

This patch needs to scan filePattern multiple times if there are more than one curly braces
in the pattern. Multilevel nested braces will lead to the number of scans close to exponential
growth.  If alternations are expanded no matter there is an embedded "/" or not, we could
use an algorithm that requires only one pass of scanning.

> File globbing alternation should be able to span path components
> ----------------------------------------------------------------
>
>                 Key: HADOOP-3498
>                 URL: https://issues.apache.org/jira/browse/HADOOP-3498
>             Project: Hadoop Core
>          Issue Type: Improvement
>          Components: fs
>            Reporter: Tom White
>            Assignee: Tom White
>         Attachments: hadoop-3498.patch
>
>
> For example, {/a/b,/c/d} should expand to /a/b and /c/d. This change would also permit
a consistent syntax for specifying multiple input paths to MapReduce, streaming and Pig by
specification of a single glob path with alternation {/a/b,/c/d}, rather than a collection
of comma separated glob paths /a/b,/c/d.
> This change would also make globbing more consistent with bash, which supports this feature.

-- 
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