[
https://issues.apache.org/jira/browse/CASSANDRA-5675?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Jonathan Ellis updated CASSANDRA-5675:
--------------------------------------
Fix Version/s: 2.0 beta 1
> cqlsh shouldn't display "null" for empty values
> -----------------------------------------------
>
> Key: CASSANDRA-5675
> URL: https://issues.apache.org/jira/browse/CASSANDRA-5675
> Project: Cassandra
> Issue Type: Bug
> Reporter: Sylvain Lebresne
> Assignee: Aleksey Yeschenko
> Priority: Minor
> Labels: cqlsh
> Fix For: 1.2.7, 2.0 beta 1
>
> Attachments: 5675.txt
>
>
> For historical reason (and compatibility with thrift), all type support an
> empty value, even type like int for which it doesn't really make sense (see
> CASSANDRA-5674 too on that subject).
> If you input such an empty value for a type like int, cqlsh will display it
> as null:
> {noformat}
> cqlsh:ks> CREATE TABLE test (k text PRIMARY KEY, v int);
> cqlsh:ks> INSERT INTO test(k, v) VALUES ('someKey', blobAsInt(0x));
> cqlsh:ks> SELECT * FROM test;
> k | v
> ---------+------
> someKey | null
> {noformat}
> But that's not correct, it suggests {{v}} has no value but that's not true,
> it has a value, it's just an empty one.
> Now, one may argue support empty values for a type like int is broken, and I
> would agree with that. But thrift allows it so we probably need to preserve
> that behavior for compatibility sake. And I guess the need to use blobAsInt
> at least make it clear that it's kind of a hack.
> That being said, cqlsh should not display null as this is confusing. Instead
> I'd suggest either displaying nothing (that's how an empty string is
> displayed after all), or to just go with some random explicit syntax like say
> "[empty value]"
--
This message was sent by Atlassian JIRA
(v6.2#6252)
