accumulo-notifications mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Keith Turner (JIRA)" <>
Subject [jira] [Commented] (ACCUMULO-1346) deleterows should take a -r option
Date Thu, 18 Jul 2013 18:08:49 GMT


Keith Turner commented on ACCUMULO-1346:

bq. Except in a sharded data style you wouldn't be getting rid of the tablet like you would
with the deleteRow() option.

Good point, I did not think of that case (where deleting a single row creates an empty tablet)
w/ the compaction option.  To get that behavior would need to merge if empty following the
compaction.  This could still be done in the shell code.  Or we could add the deleteRow call
to the API.  I was just thinking it would become redundant if inclusiveness were supported
in the API in the future, I am not opposed to that option.  An efficient implementation for
-r would be useful in the shell now, and could be easily achieved.   I think properly making
the API efficiently support inclusive start row would be more time consuming.   I think [~cmccubbin]'s
suggestion would work, but it seems like it could cause two compactions of the same data.

> deleterows should take a -r option
> ----------------------------------
>                 Key: ACCUMULO-1346
>                 URL:
>             Project: Accumulo
>          Issue Type: Improvement
>          Components: shell
>    Affects Versions: 1.4.2
>            Reporter: Mike Drob
>            Assignee: Drew Farris
>            Priority: Minor
>              Labels: hackathon, hackation-5-2013, newbie
> deleterows only allows an exclusive begin row and an inclusive end row. This is difficult
to use for deleting a single (large) row. It would be nice if a user could specify -r, similar
to scan.

This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see:

View raw message