cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Sylvain Lebresne (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (CASSANDRA-2321) Counter column values shows in hex values. Need to show it in string value.
Date Fri, 25 Mar 2011 19:02:05 GMT

    [ https://issues.apache.org/jira/browse/CASSANDRA-2321?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13011356#comment-13011356
] 

Sylvain Lebresne commented on CASSANDRA-2321:
---------------------------------------------

bq. I'm not sure this refactor is much of an improvement. Seems like the code moves around
a lot but volume and complexity are not really reduced.

I do believe the refactor is an improvement (granted, maybe not a huge one). During validation
we do a bunch of queries to DatabaseDescriptor to check the cf exists, then to get its type
(super or standard), then to get its value validator, etc... Granted those are just hashMaps
gets, but they are just unnecessary. I do think that the method used by the refactor, that
is getting the metadata once and giving it to all other method is cleaner. There also the
fact that we revalidate the column family for each Mutation of a batch_mutate, even if all
of them are on the same cf, which this refactor fixes too.
I don't care so much about those and won't fight over it, but I think it would be a pity to
leave ThriftValidation in that state (and the refactoring is really trivial). Note that I'll
be perfectly ok with moving the refactor in another ticket if that makes it better.

bq. Can't we just add a "validateCounterCF" to counter calls and "validateNonCounterCF" otherwise?

Yes we could, up to the fact that the refactor would still make sense I think for the reason
above :)

bq. (I'd prefer to say "counter/noncounter" vs "noncommutative/commutative" but if you really
prefer the other that's okay too.)

When I have a few free cycles I plan to remove the 'commutative' wording completely as I think
this was premature generalization and make the code less readable right now. So I'm on your
side, I just went for consistency with the rest of the code for now.

> Counter column values shows in hex values. Need to show it in string value.
> ---------------------------------------------------------------------------
>
>                 Key: CASSANDRA-2321
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-2321
>             Project: Cassandra
>          Issue Type: Bug
>          Components: Core
>    Affects Versions: 0.8
>         Environment: Linux
>            Reporter: Mubarak Seyed
>            Assignee: Sylvain Lebresne
>            Priority: Minor
>             Fix For: 0.8
>
>         Attachments: 0001-Don-t-allow-normal-query-on-counter-CF.patch
>
>
> CounterColumnType.getString() returns hexString.
> {code}
> public String getString(ByteBuffer bytes)
> { 
>        return ByteBufferUtil.bytesToHex(bytes);
> }
> {code}
> and python stress.py reader returns
> [ColumnOrSuperColumn(column=None, super_column=SuperColumn(name='19', columns=[Column(timestamp=1299984960277,
name='56', value='\x7f\x00\x00\x01\x00\x00\x00\x00\x00\x00\x00\x08\x00\x00\x00\x00\x00\x00\x00,',
ttl=None), Column(timestamp=1299985019923, name='57', value='\x7f\x00\x00\x01\x00\x00\x00\x00\x00\x00\x00;\x00\x00\x00\x00\x00\x00\x08\xfd',
ttl=None))]

--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira

Mime
View raw message