[
https://issues.apache.org/jira/browse/DERBY-4430?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12774776#action_12774776
]
Knut Anders Hatlen commented on DERBY-4430:
-------------------------------------------
Thanks for the patch, Sylvain! I had a quick look at it and it looks good. I'll
give it a try.
To your questions:
1) I wouldn't worry too much about the calls to getMetaData(). Both the client
driver and the embedded driver cache the DatabaseMetaData object in the
Connection object, and DatabaseMetaData caches the values returned by
stores*CaseIdentifiers, so I think there would be little to gain by caching
them ourselves in ij.
2) The ij7.sql test is run by the ToolScripts JUnit test. I see the same
failure as you if I run it with the old style runner,
o.a...functionTests.harness.RunTest, but not if I run it using the JUnit runner:
java junit.textui.TestRunner
org.apache.derbyTesting.functionTests.tests.tools.ToolScripts
3) The English locale is used to prevent the JVM's default locale from
influencing how the identifiers are handled. For instance, in Turkish locale,
toUpper("insert") returns "İNSERT" (note the upper-case i with a dot), and
toLower("INSERT") returns "ınsert" (dot-less lower-case i).
> Make ij's SHOW and DESCRIBE commands more db agnostic
> -----------------------------------------------------
>
> Key: DERBY-4430
> URL: https://issues.apache.org/jira/browse/DERBY-4430
> Project: Derby
> Issue Type: Improvement
> Components: Tools
> Affects Versions: 10.6.0.0
> Reporter: Knut Anders Hatlen
> Assignee: Sylvain Leroux
> Priority: Minor
> Attachments: DERBY-4430.ij.out, DERBY-4430.patch, DERBY-4430.sql
>
>
> ij's SHOW and DESCRIBE commands use DatabaseMetaData in order to be portable
> to other databases than Derby, and in many cases they work fine with other
> databases. However, the variants of the commands that take a table name or a
> schema name assume that unquoted identifiers are converted to and stored in
> upper case internally. This is not true for all databases, and since these
> commands don't accept quoted identifiers, there's currently no way to access
> tables/schemas that are not all upper case.
> One possible fix is to make the identifier() method in ij.jj use the
> DatabaseMetaData methods storesLowerCaseIdentifiers(),
> storesMixedCaseIdentifiers() and storesUpperCaseIdentifiers() to decide
> whether it should convert the identifier to lower case, keep it unchanged, or
> convert it to upper case. Currently, that method always converts the
> identifiers to upper case.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
