incubator-cassandra-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Machiel Groeneveld <machi...@gmail.com>
Subject Re: Trying to understand why tombstone is being read
Date Thu, 13 Feb 2014 17:43:51 GMT
We had a similar question, a column inserted with a null value also results in a tombstone.

Also see the (closed) issue about this:
https://issues.apache.org/jira/browse/CASSANDRA-6674

On 13 Feb 2014, at 16:35, Adam Patacchiola <adam@2fours.com> wrote:

> I have a table defined in cql like this:
> 
> CREATE TABLE chatmessages (
>   username    varchar,
>   spotname    varchar,
>   id          varint,
>   iv          varchar,
>   fromuser        varchar,
>   fromversion varchar,
>   touser          varchar,
>   toversion   varchar,
>   datetime    timestamp,
>   data        varchar,
>   mimetype    varchar,
>   datasize    int,
>   shareable   boolean,
>   PRIMARY KEY (username, spotname, id)
> )
> 
> That I insert messages into. Occasionally "rows" are deleted. My question is why, when
I query for a "row" using a select query specific to only that "row" which contains cells
that have never been deleted such as:
> 
>   select * from chatmessages where username='adam' and spotname='adam:cherie' and id
=1860;
> 
> Am I seeing in the cqlsh trace that 1 tombstoned cell has been read:
> 
> "Read 1 live and 1 tombstoned cells"
> 
> Thanks for any way you can help increase my understanding of what's going on here.
> 
> Adam


Mime
View raw message