[
https://issues.apache.org/jira/browse/DERBY-3886?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12635380#action_12635380
]
Rick Hillegas commented on DERBY-3886:
--------------------------------------
Hey Dag, this is pretty slick. Some small questions:
1) Does the VTI need to implement VTICosting? I think that
getEstimatedCostPerInstantiation() is probably returning a number that isn't
tuned to any particular machine. I don't think that the optimizer has a lot of
choice here and it doesn't seem to me that the VTICosting implementation will
help the optimizer pick a better plan.
2) Will the closure iterator return the roles in some reasonable sort order?
For someone just selecting from the VTI it wouldn't matter because they could
just slap an ORDER BY on the SELECT. But once the SELECT is hidden inside a
system procedure, the user can't inject an ORDER BY clause.
These are just nits. Looks like a great approach.
> SQL roles: show enabled roles
> -----------------------------
>
> Key: DERBY-3886
> URL: https://issues.apache.org/jira/browse/DERBY-3886
> Project: Derby
> Issue Type: Improvement
> Components: SQL, Tools
> Reporter: Dag H. Wanvik
> Attachments: enabled-roles.diff, enabled-roles.stat
>
>
> Rick suggested it would be nice to have something to allow inspection of
> enabled roles, cf the
> ENABLED_ROLES view of the information schema, SQL 2003, vol 11, section 5.29.
> I enclose patch suggestion for doing this via a new VTI table and syntactic
> sugar in ij:
> SYSCS_DIAG.ENABLED_ROLES and and ij command "show enabled_roles".
> Not for commit at this point, we may want to handle to general question of
> information schemata
> in another way. What say thee?
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
