hadoop-common-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Colin Patrick McCabe (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HADOOP-10987) Provide an iterator-based listing API for FileSystem
Date Thu, 02 Oct 2014 19:12:35 GMT

    [ https://issues.apache.org/jira/browse/HADOOP-10987?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14157003#comment-14157003

Colin Patrick McCabe commented on HADOOP-10987:

    // Each file system implementation should override this method and 
    // provide more efficient implementation, if possible.

Hmm, this might look better in the JavaDoc header.

+  public RemoteIterator<FileStatus> listStatusIterator(final Path p)
+  throws FileNotFoundException, IOException {
+    return listStatusIterator(p, DEFAULT_FILTER);
+  }

Should this method be final?  It seems like we'll always want it to act like the {{listStatusIterator}}
method with {{DEFAULT_FILTER}}.

-  protected RemoteIterator<LocatedFileStatus> listLocatedStatus(final Path f,
+  public RemoteIterator<LocatedFileStatus> listLocatedStatus(final Path f,
       final PathFilter filter)
I like the fact that you changed more of our subclasses to implement the two-argument form
of {{listLocatedStatus}}.  They should definitely do this when possible (like FilterFileSystem
should pass the two-argument form to what it contains).

But it seems like changing the visibility from {{protected}} to {{public}} will be a slight
incompatibility for some out-of-tree filesystems.  If they previously overrode the protected
two-argument form with a protected function, they will get a compile error now: "cannot reduce
visibility of inherited method."  Maybe we can just leave the two-argument form of {{listLocatedStatus}}
as {{protected}} for now, and handle it in a follow-on JIRA?  Just adding {{listStatusIterator}}
is already a pretty nice change.

> Provide an iterator-based listing API for FileSystem
> ----------------------------------------------------
>                 Key: HADOOP-10987
>                 URL: https://issues.apache.org/jira/browse/HADOOP-10987
>             Project: Hadoop Common
>          Issue Type: Improvement
>            Reporter: Kihwal Lee
>            Assignee: Kihwal Lee
>         Attachments: HADOOP-10987.patch, HADOOP-10987.v2.patch, HADOOP-10987.v3.patch
> Iterator based listing methods already exist in {{FileContext}} for both simple listing
and listing with locations. However, {{FileSystem}} lacks the former.  From what I understand,
it wasn't added to {{FileSystem}} because it was believed to be phased out soon. Since {{FileSystem}}
is very well alive today and new features are getting added frequently, I propose adding an
iterator based {{listStatus}} method. As for the name of the new method, we can use the same
name used in {{FileContext}} : {{listStatusIterator()}}.
> It will be particularly useful when listing giant directories. Without this, the client
has to build up a huge data structure and hold it in memory. We've seen client JVMs running
out of memory because of this.
> Once this change is made, we can modify FsShell, etc. in followup jiras.

This message was sent by Atlassian JIRA

View raw message