[ 
https://issues.apache.org/jira/browse/CASSANDRA-6870?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13941899#comment-13941899
 ] 

Edward Capriolo commented on CASSANDRA-6870:
--------------------------------------------

{quote}
this should support normal CL conventions. So if I want to trim without paxos 
because I'm confident that there are no races, I shouldn't have to pay that 
extra latency cost. I think this means we open up CL.SERIAL for normal UPDATE 
(without IF) when UDF are being invoked against data from an existing row.
{quote}
I was wondering if it made sense to support transform and transform_atomic for 
this case. IE the default transform does not have to happen in paxos at all. 

{quote}
sometimes a single output column is not enough. how do we support more? We 
could use (c1, c2) = f(c3, c4) syntax similar to what we introduced for 
composite partition keys. Or we could add "out" parameters which is what most 
SQL UDF implementations do. NB: I am fine with postponing implementation of 
this to another ticket, but I do think we should settle on the syntax ahead of 
time so that we don't accidentally make extending it harder than necessary.
{quote}

In hive UDF returns a single value. This is a pure definition of function. 
There is a different interface UDTF that returns multiple columns or even rows. 
It looks like this:
{code}
select parsurl(url, 'host', 'port') as a,b
{code}

I would suggest to differentiate a UDF which works like a mathamatical 
function,from a UDTF. Trying to make them one uber-interface would probably 
make said interface hard to implement. 


> Transform operation
> -------------------
>
>                 Key: CASSANDRA-6870
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-6870
>             Project: Cassandra
>          Issue Type: New Feature
>            Reporter: Edward Capriolo
>            Assignee: Edward Capriolo
>            Priority: Minor
>
> Compare and swap uses paxos to only update a value only if some criteria is 
> met. If I understand correctly we should be able to use this feature to 
> provide a wider variety of server side operations. 
> For example inside a paxos transaction performing a slice and then using a 
> function to manipulate the slice. You could accomplish features like append 
> and increment this way without user needing to know the current value.
> I took a stab at doing this. I **think** I did it correctly. Comments welcome.



--
This message was sent by Atlassian JIRA
(v6.2#6252)

Reply via email to