cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Le Duc Duy (JIRA)" <>
Subject [jira] [Commented] (CASSANDRA-5851) Fix 2i on composite components omissions
Date Wed, 07 Aug 2013 05:27:48 GMT


Le Duc Duy commented on CASSANDRA-5851:

I got this when I query on 2 indexed column. Is this expected?

SELECT * FROM gurugara.outgoing_edge WHERE source_id = 3 AND dest_id = 1 AND type = 'meh'
`dest_id` and `type` are secondary indexed column

ERROR 07:23:13,702 Exception in thread Thread[ReadStage:4,5,main]
java.lang.RuntimeException: java.lang.RuntimeException: Unable to search across multiple secondary
index types
	at org.apache.cassandra.service.StorageProxy$
	at java.util.concurrent.ThreadPoolExecutor.runWorker(
	at java.util.concurrent.ThreadPoolExecutor$
Caused by: java.lang.RuntimeException: Unable to search across multiple secondary index types
	at org.apache.cassandra.db.RangeSliceCommand.executeLocally(
	at org.apache.cassandra.service.StorageProxy$LocalRangeSliceRunnable.runMayThrow(
	at org.apache.cassandra.service.StorageProxy$
	... 3 more

> Fix 2i on composite components omissions
> ----------------------------------------
>                 Key: CASSANDRA-5851
>                 URL:
>             Project: Cassandra
>          Issue Type: Bug
>    Affects Versions: 2.0 beta 2
>            Reporter: Aleksey Yeschenko
>            Assignee: Aleksey Yeschenko
>             Fix For: 2.0
>         Attachments: 5851.txt
> There some edge-cases, not covered by CASSANDRA-5125, the attached patch fixes those:
> (Assuming CREATE TABLE test (pk0 int, pk1 int, ck0 int, ck1 int, val int, PRIMARY KEY
((pk0, pk1), ck0, ck1)))
> - could not create a 2i on the first part of a composite partition key (pk0)
> - if created, it couldn't work because of getKeyBounds() returning non-empty bounds
> - could create an index on the first clustering key column (ck0), but it would never
actually be triggered on reads
> - queries like SELECT * FROM test WHERE pk0 = x AND pk1 = y AND ck1 = z would throw an
exception because COCK.makeIndexColumnNameBuilder() couldn't handle empty provided columnName
> - cqlsh could not describe any of these indexes because it was taking column aliases
and key aliases from schema_columnfamilies and not reading them directly from schema_columns
(had to do the related refactoring).

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