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-8374) Better support of null for UDF
Date Mon, 15 Dec 2014 11:00:14 GMT

    [ https://issues.apache.org/jira/browse/CASSANDRA-8374?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14246553#comment-14246553
] 

Sylvain Lebresne commented on CASSANDRA-8374:
---------------------------------------------

The thing is, if you have a simple function like
{noformat}
CREATE FUNCTION incr(x int) RETURNS int LANGUAGE JAVA AS 'return x + 1';
{noformat}
then I'm not sure it's very friendly for
{noformat}
SELECT a, b, incr(c) FROM foo;
{noformat}
to crash when it encounter a row with no value for {{c}}. Especially since that would be throw
in the middle of the execution and we don't even have a good exception to return for that.
Besides, since a {{null}} means "there is no value", it makes a lot of sense to me that the
function is simply not applied and "there is no value" is returned. In fact, I'm kind of -1
for throwing an exception in that particular case, because imho that would make UDT without
{{ALLOW NULLS}} largely unusable in practice.

Of course, when you have
{noformat}
SELECT * FROM foo WHERE v = incr(null);
{noformat}
then I can agree that that throwing an IRE is an option. But I wonder if handling {{null}}
differently in those two cases won't confuse people more than if we have a single rule that
is applied everywhere. Typically in something like:
{noformat}
SELECT x, add(a, ?) FROM foo;
{noformat}
it feels more consistent to have the same behavior whether {{a}} or the bind marker is {{null}}.

bq. handling the combination of UDA + non-ALLOW NULLS state function (thinking about what
happens when a ALLOW NULLS UDF is replaced by a non-ALLOW NULLS variant)

I don't think we should allow such replacement. We should allow to replace a function only
if it has the exact same signature *and* the same handling of nulls or it's going to be messy.

> Better support of null for UDF
> ------------------------------
>
>                 Key: CASSANDRA-8374
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-8374
>             Project: Cassandra
>          Issue Type: Bug
>            Reporter: Sylvain Lebresne
>            Assignee: Robert Stupp
>             Fix For: 3.0
>
>         Attachments: 8473-1.txt
>
>
> Currently, every function needs to deal with it's argument potentially being {{null}}.
There is very many case where that's just annoying, users should be able to define a function
like:
> {noformat}
> CREATE FUNCTION addTwo(val int) RETURNS int LANGUAGE JAVA AS 'return val + 2;'
> {noformat}
> without having this crashing as soon as a column it's applied to doesn't a value for
some rows (I'll note that this definition apparently cannot be compiled currently, which should
be looked into).  
> In fact, I think that by default methods shouldn't have to care about {{null}} values:
if the value is {{null}}, we should not call the method at all and return {{null}}. There
is still methods that may explicitely want to handle {{null}} (to return a default value for
instance), so maybe we can add an {{ALLOW NULLS}} to the creation syntax.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message