cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jonathan Ellis (JIRA)" <j...@apache.org>
Subject [jira] Commented: (CASSANDRA-1861) multiget_slice does not return correct columns based on column slice
Date Tue, 14 Dec 2010 18:25:00 GMT

    [ https://issues.apache.org/jira/browse/CASSANDRA-1861?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12971354#action_12971354
] 

Jonathan Ellis commented on CASSANDRA-1861:
-------------------------------------------

you'd need to page through the columns client-side until column.startswith(X) is no longer
true

> multiget_slice does not return correct columns based on column slice
> --------------------------------------------------------------------
>
>                 Key: CASSANDRA-1861
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-1861
>             Project: Cassandra
>          Issue Type: Bug
>          Components: Core
>    Affects Versions: 0.7.0 rc 2
>            Reporter: chirag patel
>            Priority: Critical
>
> We are using Random Partitioner.
> We are trying to do a multiget_slice on a column family which is sorted by UTF8Type and
it does not behave as expected.
> ColumnFamily: RecordIndexes
>       Columns sorted by: org.apache.cassandra.db.marshal.UTF8Type
>       Row cache size / save period: 0.0/0
>       Key cache size / save period: 200000.0/3600
>       Memtable thresholds: 0.29062499999999997/62/60
>       GC grace seconds: 864000
>       Compaction min/max thresholds: 4/32
>       Read repair chance: 1.0
> So RECORDINDEXES CF has 1 key with 3 columns as shown below.
> In [6]: RECORDINDEXES.get('testkey')
> Out[6]: {u'A_1.2.3.4_1': '1', u'N_edited.test.com_1': '1', u'P_45653_1': '1'}
> When I do a multiget_slice with column_start as "N_edi.test.com._" and finish column
as 'P_', I should not get any thing as there is no column starting with "N_edi.test.com._"
as column name, but I do get one column back which is incorrect.
> ***********************************
> In [11]: RECORDINDEXES.get('testkey', column_start='N_edi.test.com._',column_finish='P_')
> Out[11]: {u'N_edited.test.com_1': '1'}
> ***********************************
> Some more invalid results
> ***********************************
> In [9]: RECORDINDEXES.get('testkey', column_start='NS_',column_finish='P_')
> Out[9]: {u'N_edited.test.com_1': '1'}
> In [10]: 
> In [16]: RECORDINDEXES.get('testkey', column_start='NS',column_finish='P_')
> Out[16]: {u'N_edited.test.com_1': '1'}
> In [17]: 
> ***********************************
> I tried this with hector client library as well, so its the same result.
> Can anybody please look into this?
> This is blocking issue for us as our data retrieval heavily depends on multiget_slice
query.
> FYI: beta3 release also had this problem, so we upgraded to rc2 and still seeing the
same problem.
> Thanks,
> -Chirag. 

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


Mime
View raw message