[
https://issues.apache.org/jira/browse/DERBY-6904?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16027844#comment-16027844
]
Bryan Pendleton commented on DERBY-6904:
----------------------------------------
I spent some time studying your most recent patch, and it looks great! It
applies and builds cleanly, and all the existing tests pass for me.
Question: it seems like the new AUTOINCREMENTCYCLE column in SYSCOLUMNS is of
BIGINT type? Doesn't it simply take the values 1 or 0? It seems like we should
make this column BOOLEAN or (perhaps) SMALLINT? Or was there a reason that we
needed to use BIGINT here?
Question: can you remind me what we decided about the SYSSEQUENCES question
that you raised on March 13th? I think we decided that we are not going to
modify anything about SYSSEQUENCES, but we didn't seem to document that
decision or its reasoning in this JIRA?
Also, a little nit: the change to IdentitySequenceTest in the patch looks
incomplete. Instead of simply commenting out the test which has (expected)
changed behavior after this patch, we should update the comments and modify the
test code at line 690 so that it verifies the new expected behavior.
Meanwhile, I'm trying to understand what work is left here (besides the data
type question above): is it: (1) write new test cases, and (2) write new
documentation?
Perhaps you and I can split up some of this remaining work; also, perhaps we
can handle the test cases and documentation as separate sub-tasks of this JIRA,
to keep it organized for reviewers who are studying the work?
For example, we could (perhaps) compile a simple numbered list of all the new
test cases that we want to write, and then you could take some of them and I
could take some of them and we could merge our new test cases together later?
Also, I think it would be fine to add any documentation as a separate sub-task,
and we can certainly commit the code+tests prior to committing the
documentation.
I think that's all I have for now. Thanks again for the great work on this
feature, it's much appreciated!
> Enhance ALTER TABLE to allow CYCLE option to be changed.
> --------------------------------------------------------
>
> Key: DERBY-6904
> URL: https://issues.apache.org/jira/browse/DERBY-6904
> Project: Derby
> Issue Type: Improvement
> Components: SQL
> Affects Versions: 10.13.1.0
> Reporter: Bryan Pendleton
> Assignee: Harshvardhan Gupta
> Priority: Minor
> Attachments: 6904_2.diff, 6904_3.diff, 6904.diff, columnMismatch,
> compareDbName
>
>
> It would be nice to have a variant of the ALTER TABLE statement,
> similar to the SET INCREMENT BY or SET RESTART WITH variants,
> that allowed a user to change the CYCLE option on an IDENTITY column.
> I think that the code flow for this could be closely modelled on the
> code flow for the SET INCREMENT BY variant: after the statement
> is parsed, we simply need to drop and recreate the underlying SEQUENCE
> object with the desired CYCLE option (and preserving all the other
> aspects of the underlying SEQUENCE object.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
