[
https://issues.apache.org/jira/browse/DRILL-2463?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14365904#comment-14365904
]
Daniel Barclay (Drill) commented on DRILL-2463:
-----------------------------------------------
Current attempts to fix the lower layer now seem to result in a problem with
scalar replacement manipulation of byte code.
To avoid mixing in below-JDBC cleanup/soft-bug changes with JDBC hard-bug*
changes and delaying the JDBC bug-fix changes until lower-level problems and
indirect requirements are understood and solved (to avoid the fate of
DRILL-1735--having otherwise-independent changes delayed by other things), I
think we should implement all the checks in AvaticaDrillSqlAccessor now, filing
a Jira report and putting TODO notes in AvaticaDrillSqlAccessor to later
implement the non-primitive-type checks in the lower layer (and then remove the
then-redundant non-primitive-type checks from AvaticaDrillSqlAccessor).
(*Calling ResultSet.getBoolean(...) when the value is SQL NULL throws an
exception (and ResultSet.isNull() can't be used without first calling a
getXxx(...) method for the column).)
> Implement JDBC mapping of SQL NULL for ResultSet.getXxx() methods
> -----------------------------------------------------------------
>
> Key: DRILL-2463
> URL: https://issues.apache.org/jira/browse/DRILL-2463
> Project: Apache Drill
> Issue Type: Bug
> Reporter: Daniel Barclay (Drill)
> Assignee: Daniel Barclay (Drill)
>
> Fix AvaticaDrillSqlAccessor to implement mapping of SQL NULL to dummy
> primitive values (e.g,., returning 0 for ResultSet.getInt(...)).
> Fix SqlAccessors template to implement mapping of SQL NULL to null pointers
> (e.g., returning null from ResultSet.getString(...).)
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
