Oracle TimesTen Dialect for Hibernate 6.6 Updates

[carlblan]

Some updates from an Oracle employee to the Oracle TimesTen Community dialect. ( Hibernate 6.6 )

In TimesTenDialect.java

Adding more datatypes into columnType()
Added more SQL functions support into initializeFunctionRegistry()
Updated the custom definition for 'getForUpdateString()'
Added a custom definition for 'getNativeIdentifierGeneratorStrategy()'
Added a custom definition for 'currentDate()'
Added a custom definition for 'currentTime()'
Added a custom definition for 'getMaxVarcharLength()'
Added a custom definition for 'getMaxVarbinaryLength()'
Added a custom definition for 'isEmptyStringTreatedAsNull()'
Added a custom definition for 'supportsTupleDistinctCounts()'
Added a custom definition for 'getDual()'
Added a custom definition for 'getFromDualForSelectOnly()'

In TimesTenSqlAstTranslator.java

Added a custom definition for 'renderRowsToClause()'

In TimesTenLimitHandler.java

    The class now extends 'AbstractLimitHandler'
    Removed a custom definition for 'insert()'
    Added a custom definition for 'supportsLimit()'
    Added a custom definition for 'supportsOffset()'
    Added a custom definition for 'supportsLimitOffset()'
    Added a custom definition for 'supportsVariableLimit()'
    Added a custom definition for 'convertToFirstRowValue(int zeroBasedFirstResult)'
    Added a custom definition for 'useMaxForLimit()'
    Added a custom definition for 'limitClause(boolean hasFirstRow)'

In TimesTenSequenceSupport.java

    The Class now implements 'SequenceSupport'
    Added a custom definition for 'supportsSequences()'
    Added a custom definition for 'supportsPooledSequences()'
    Added a custom definition for 'getSelectSequenceNextValString(String sequenceName)'
    Added a custom definition for 'getSequenceNextValString(String sequenceName)'
    Added a custom definition for 'getCreateSequenceString(String sequenceName)'
    Added a custom definition for 'getDropSequenceString(String sequenceName)'

Testing:

The changes were tested with some local tests we have using this Hibernate 6 TimesTen community dialect.

---

By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license
and can be relicensed under the terms of the LGPL v2.1 license in the future at the maintainers' discretion.
For more information on licensing, please check here.

---

[hibernate-github-bot]

Thanks for your pull request!

This pull request does not follow the contribution rules. Could you have a look?

❌ All commit messages should start with a JIRA issue key matching pattern HHH-\d+
    ↳ Offending commits: [e6f5c1e, 27e72ac, 0e517d9, f859b93, fb08af3, 8a8354a, 894124a, dfdfbcb, 8658705, fe8b24d, 197ee03]

› This message was automatically generated.

[carlblan]

@beikov @mbellade @yrodiere

Here's a new PR to update the TimesTen Community Dialect for Hibernate 6.6

I closed the previous PR: #10273
Since I accidentally fetched/updated from main instead of 6.6 branch

This is a clean PR with the same changes but addressed the previous comments:

• Removed the copyright headers from my changes since Oracle has been added as a corporate contributor.
• Addressed comments for 'initializeFunctionRegistry()'

Let me know any other comment.

Thanks

[beikov]

Does TimesTen not support a parameter here?

Suggested change

	renderFetchPlusOffsetExpressionAsLiteral( fetchClauseExpression, offsetClauseExpression, -1 );
	renderFetchPlusOffsetExpressionAsSingleParameter( fetchClauseExpression, offsetClauseExpression, -1 );

[carlblan]

I've tried using 'renderFetchPlusOffsetExpressionAsSingleParameter()'. But it doesn't work as expected, that's why I ended using the 'AsLiteral' version.

Let me explain...

It seems like if I use:
'renderFetchPlusOffsetExpressionAsSingleParameter(
Expression fetchClauseExpression,
Expression offsetClauseExpression,
int offset
)'

The third parameter 'int offset' which is set to '-1' is not being added to the calculation. The end resultSet has 'rowsLimit +1 row'.

But If I use 'renderFetchPlusOffsetExpressionAsLiteral()' with the 'int offset=-1' my end resultSet has exactly the 'rowsLimit' size. Which is what we want.

[beikov]

Looking at the implementation of AbstractSqlAstTranslator#renderFetchPlusOffsetExpressionAsSingleParameter, it should work. It might be the case that a fix that I did for HHH-19632 needs to be backported to ORM 6.6 to make this work fully. Can you please share how you were testing/verifying this?

[carlblan]

I see, we are testing it like:

      session.beginTransaction();
      int rowOffset    = 27;
      int maxRowsLimit = 93;

      List<Book> books = session.createQuery("FROM Book ORDER BY id", Book.class)
        .setFirstResult(rowOffset)   // Offset
        .setMaxResults(maxRowsLimit) // Limit
        .list();

      // Check max rows limit
      Assertions.assertEquals( maxRowsLimit, books.size() );

And this fails with:
[ERROR] TestLimitHandler.testPaginationQueryOffsetAndLimit:136 expected: <93> but was: <94>

Note: the table contains 300 rows we expected to have 93 books (MaxRowsLimit)

[beikov]

Please try to reuse the existing CommonFunctionFactory methods as much as possible, since they also contain parameter type inference information and other extras that you are missing right now.

[carlblan]

Hi @beikov , Could you help with a new review for this PR.

Thank you in advance.

[beikov]

Hi @beikov , Could you help with a new review for this PR.

Thank you in advance.

Hi, there are still a few comments that I made which you did not address. When that is done, I can give it another look.

[carlblan]

Hi @beikov I just updated the PR.
For the remaining comments I added an explanation on why those are still not addressed.
Could you help me reading/checking those comments I made?

Thank you in advance.

[beikov]

I don't see any comments or changes that address the my comments @carlblan, so maybe you forgot to press send on them? Or forgot to push the correct changes?

[carlblan]

Hi @beikov,
I addressed most of the comments, but for the remaining ones (4) I added an explanation as a reply for each comment.

And again, thanks for your time helping with the review.

Kind regards

[carlblan]

I've tried using 'renderFetchPlusOffsetExpressionAsSingleParameter()'. But it doesn't work as expected, that's why I ended using the 'AsLiteral' version.

Let me explain...

It seems like if I use:
'renderFetchPlusOffsetExpressionAsSingleParameter(
Expression fetchClauseExpression,
Expression offsetClauseExpression,
int offset
)'

The third parameter 'int offset' which is set to '-1' is not being added to the calculation. The end resultSet has 'rowsLimit +1 row'.

But If I use 'renderFetchPlusOffsetExpressionAsLiteral()' with the 'int offset=-1' my end resultSet has exactly the 'rowsLimit' size. Which is what we want.

[carlblan]

Changes:

• Addressing latest comments in initializeFunctionRegistry() : For 'getdate', 'trunc' and 'round'
• Answering question regarding the use of renderFetchPlusOffsetExpressionAsSingleParameter()

Let me know if anything.
Thanks.

[yrodiere]

@carlblan I approved the workflow run to let you see if everything works and address any problems, but Christian is currently away, back in two weeks. If you think this needs to get in urgently, please reach out on Zulip and maybe someone with the required expertise will be able to take a look at your latest changes.

[carlblan]

@carlblan I approved the workflow run to let you see if everything works and address any problems, but Christian is currently away, back in two weeks. If you think this needs to get in urgently, please reach out on Zulip and maybe someone with the required expertise will be able to take a look at your latest changes.

@yrodiere Got it. No problem, I can wait for Christian.
Thanks.

[yrodiere]

Thanks for your understanding. I the meantime, it seems tests are failing 😬

[beikov]

I found a bug in the handling of static offsets which I added a fix for and also cleaned up your PRs commits. This is superseded by #11140

Can you please test that PR @carlblan and tell me on the other PR comments if that works out for you?

[carlblan]

Hi @beikov I just tested your new changes and replied as commend in #11140

Thanks