get_range_slices problem with super columns

[Jianing Hu <ji...@gmail.com>]

I'm seeing cases where the count in slicerange predicate is not
respected. This is only happening for super columns. I'm running
Cassandra 0.6.4 in a single node.

Steps to reproduce, using the Keyspace1.Super1 CF:
* insert three super columns, bar1 bar 2, and bar3, under the same key
* delete bar1
* insert bar1 again
* run a get_range_slices on Super1, with start=bar1, finish=bar3, and count=1
* I expected only bar1 to be returned, but both both bar1 and bar2 are
returned. bar3 isn't, though. so count is somewhat respected.

I've filed a jira with a test script attached:
https://issues.apache.org/jira/browse/CASSANDRA-1591

Thanks for any advice.

- Jianing

Re: get_range_slices problem with super columns

[Jianing Hu <ji...@gmail.com>]

Was anyone able to reproduce this bug?

On Wed, Oct 6, 2010 at 6:19 PM, Jianing Hu <ji...@gmail.com> wrote:
> I'm seeing cases where the count in slicerange predicate is not
> respected. This is only happening for super columns. I'm running
> Cassandra 0.6.4 in a single node.
>
> Steps to reproduce, using the Keyspace1.Super1 CF:
> * insert three super columns, bar1 bar 2, and bar3, under the same key
> * delete bar1
> * insert bar1 again
> * run a get_range_slices on Super1, with start=bar1, finish=bar3, and count=1
> * I expected only bar1 to be returned, but both both bar1 and bar2 are
> returned. bar3 isn't, though. so count is somewhat respected.
>
> I've filed a jira with a test script attached:
> https://issues.apache.org/jira/browse/CASSANDRA-1591
>
> Thanks for any advice.
>
> - Jianing
>