You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@cassandra.apache.org by "Edward Capriolo (JIRA)" <ji...@apache.org> on 2010/06/21 18:49:23 UTC
[jira] Created: (CASSANDRA-1215) Server Side Operations
Server Side Operations
----------------------
Key: CASSANDRA-1215
URL: https://issues.apache.org/jira/browse/CASSANDRA-1215
Project: Cassandra
Issue Type: New Feature
Components: Core
Reporter: Edward Capriolo
Cassandra values are byte arrays. To operate on these byte arrays a client will have to get the value from a server, modify it , and retransmit the value back to be set. This is NOT a request for atomic operations, however some types of atomic operations may be possible after vector clocks are implemented. Regardless of vector clocks or atomic operations, some common string operations would still be useful.
These type of functions may include:
{noformat}
append
substring
increment
indexof
{noformat}
Operations that work on lists would be more challenging again because Cassandra does not know or care what the underlying column data is, but those could be specified in the method call.
{noformat}
pop (String delimieter )
itemat(String delimeter, item i)
{noformat}
or possibly described in the schema (I do not like this idea but wanted to mention it)
{noformat}
<ColumnFamily valueDelimeter"\t"> {noformat}
Also theoretically a user could pass an object implementing an interface or a string that is a "little language" that operates on the data to return some result or change the data.
I would like to discuss the merits of such features, and if we decide these would be useful I would like to work on implementing them. I can not assign myself this ticket otherwise I would have.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (CASSANDRA-1215) Server Side Operations
Posted by "Jeff Hammerbacher (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CASSANDRA-1215?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12881003#action_12881003 ]
Jeff Hammerbacher commented on CASSANDRA-1215:
----------------------------------------------
Hey Ed,
I think the functions you propose are best implemented in a coprocessor-like framework; see CASSANDRA-1016.
Regards,
Jeff
> Server Side Operations
> ----------------------
>
> Key: CASSANDRA-1215
> URL: https://issues.apache.org/jira/browse/CASSANDRA-1215
> Project: Cassandra
> Issue Type: New Feature
> Components: Core
> Reporter: Edward Capriolo
>
> Cassandra values are byte arrays. To operate on these byte arrays a client will have to get the value from a server, modify it , and retransmit the value back to be set. This is NOT a request for atomic operations, however some types of atomic operations may be possible after vector clocks are implemented. Regardless of vector clocks or atomic operations, some common string operations would still be useful.
> These type of functions may include:
> {noformat}
> append
> substring
> increment
> indexof
> {noformat}
> Operations that work on lists would be more challenging again because Cassandra does not know or care what the underlying column data is, but those could be specified in the method call.
> {noformat}
> pop (String delimieter )
> itemat(String delimeter, item i)
> {noformat}
> or possibly described in the schema (I do not like this idea but wanted to mention it)
> {noformat}
> <ColumnFamily valueDelimeter"\t"> {noformat}
> Also theoretically a user could pass an object implementing an interface or a string that is a "little language" that operates on the data to return some result or change the data.
> I would like to discuss the merits of such features, and if we decide these would be useful I would like to work on implementing them. I can not assign myself this ticket otherwise I would have.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.