[
https://issues.apache.org/jira/browse/CASSANDRA-11398?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15206247#comment-15206247
]
Sam Tunnicliffe commented on CASSANDRA-11398:
---------------------------------------------
cc [~xedin], [~jrwest]
> Missing results when merging from SASI index sources
> ----------------------------------------------------
>
> Key: CASSANDRA-11398
> URL: https://issues.apache.org/jira/browse/CASSANDRA-11398
> Project: Cassandra
> Issue Type: Bug
> Components: sasi
> Reporter: Sam Tunnicliffe
> Fix For: 3.x
>
>
> There seems to be a problem in {{QueryPlan.ResultIterator}} or
> {{RangeIterator}} when merging results from mulitple index locations. I've
> only repro'd using a prefix index on a clustering column, but in that case it
> seems to manifest whenever the search results are drawn from > 1 index,
> either multiple sstable indexes or a combination of memtable and sstable
> indexes).
> Note, the following repro case is masked by CASSANDRA-11397, so the patch
> attached to that issue needs to be applied to expose this.
> {code}
> cqlsh> create table ks.t1 (k text, c1 text, c2 text, c3 text, v text, primary
> key (k,c1,c2,c3));
> cqlsh> create custom index on ks.t1(c2) using
> 'org.apache.cassandra.index.sasi.SASIIndex';
> cqlsh> insert into ks.t1(k, c1, c2, c3, v) values ('a', 'ba', 'ca', 'da',
> 'va');
> cqlsh> select * from ks.t1 where c1 = 'ba' and c2 like 'c%' and c3 = 'da'
> allow filtering;
> k | c1 | c2 | c3 | v
> ---+----+----+----+----
> a | ba | ca | da | va
> (1 rows)
> # flush
> cqlsh> insert into ks.t1(k, c1, c2, c3, v) values ('a', 'ba', 'ca', 'db',
> 'va');
> cqlsh> select * from ks.t1 where c1 = 'ba' and c2 like 'c%' and c3 = 'da'
> allow filtering;
> k | c1 | c2 | c3 | v
> ---+----+----+----+---
> (0 rows)
> # flush again
> cqlsh> select * from ks.t1 where c1 = 'ba' and c2 like 'c%' and c3 = 'da'
> allow filtering;
> k | c1 | c2 | c3 | v
> ---+----+----+----+---
> (0 rows)
> # compact
> cqlsh> select * from ks.t1 where c1 = 'ba' and c2 like 'c%' and c3 = 'da'
> allow filtering;
> k | c1 | c2 | c3 | v
> ---+----+----+----+----
> a | ba | ca | da | va
> (1 rows)
> {code}
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
