cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jonathan Ellis (JIRA)" <>
Subject [jira] [Commented] (CASSANDRA-3036) Vague primary key references in CQL
Date Wed, 17 Aug 2011 20:43:27 GMT


Jonathan Ellis commented on CASSANDRA-3036:

- let's name the ColumnDef variable in validateCfDef something other than "column" (which
usually means an IColumn or similar)
- when we throw an invalid column name error, we should use the CF's comparator rather than
AsciiType (which will error out on many binary column names)
- CCFS calls getByteBuffer which catches MarshallException and turns it into IRE.  Suggest
doing the validation ones from CCFS.validate, that way we don't have to do it again in CCFS.getColumns.
 (Can just call AbstractType.fromString directly)
- go ahead and commit the println removal separately :)

> Vague primary key references in CQL
> -----------------------------------
>                 Key: CASSANDRA-3036
>                 URL:
>             Project: Cassandra
>          Issue Type: Bug
>          Components: API
>    Affects Versions: 0.8.1
>            Reporter: Kelley Reynolds
>            Assignee: Pavel Yaskevich
>            Priority: Minor
>              Labels: core, cql
>             Fix For: 0.8.5
>         Attachments: CASSANDRA-3036.patch
> create columnfamily wonk (id 'utf8' primary key, id int)
> update wonk set id=1 where id='test'
> create index wonk_id on wonk (id)
> This does what you would expect but then the results are unclear when using 'id' in a
where clause.
> "select * from wonk where id=1" returns nothing and "select * from wonk where id='test'"
works fine.
> Perhaps secondary indexes should not be allowed on columns that have the same name as
the key_alias? At least a warning or something should be thrown to indicate you've just made
a useless index.

This message is automatically generated by JIRA.
For more information on JIRA, see:


View raw message