groovy-notifications mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "ASF GitHub Bot (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (GROOVY-8288) [Sql] withBatch fails when batchSize == number of addBatch call
Date Tue, 15 Aug 2017 07:55:00 GMT

    [ https://issues.apache.org/jira/browse/GROOVY-8288?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16126906#comment-16126906
] 

ASF GitHub Bot commented on GROOVY-8288:
----------------------------------------

Github user akapps commented on a diff in the pull request:

    https://github.com/apache/groovy/pull/586#discussion_r133134481
  
    --- Diff: subprojects/groovy-sql/src/main/java/groovy/sql/BatchingStatementWrapper.java
---
    @@ -83,8 +83,15 @@ public void clearBatch() throws SQLException {
         }
     
         public int[] executeBatch() throws SQLException {
    -        int[] lastResult = delegate.executeBatch();
    -        processResult(lastResult);
    +        if (batchCount > 0) {
    +            int[] lastResult = delegate.executeBatch();
    +            processResult(lastResult);
    +        }
    +        else if (results.isEmpty()) {
    +            log.warning("Nothing has been added to batch");
    +            // we let the JDBC provider decide how to handle an empty batch execution
    +            return delegate.executeBatch();
    +        }
    --- End diff --
    
    This "else if" is not really elegant for such a particular case...
    I could have done `if (batchCount > 0 || results.isEmpty())` in the first condition,
but I thought it made the intention unreadable.


> [Sql] withBatch fails when batchSize == number of addBatch call
> ---------------------------------------------------------------
>
>                 Key: GROOVY-8288
>                 URL: https://issues.apache.org/jira/browse/GROOVY-8288
>             Project: Groovy
>          Issue Type: Bug
>          Components: SQL processing
>    Affects Versions: 2.5.0-beta-1, 2.4.12
>            Reporter: Antoine Kapps
>
> {{Sql.withBatch(batchSize, ..)}} calls {{delegate.executeBatch()}} both when batchCount
reaches batchSize, and at the end of the method (after the call to the "batch statement" closure).
> Thus, if the number of {{addBatch()}} calls in the closure is exactly the same as, or
a multiple of {{batchSize}}, one too much call to {{executeBatch()}} is made : nothing has
been added to the batch.
> Which, at least with HSQLDB 2.4.0, leads to a SQLException on that very last call.
> Test to reproduce (compliant with SqlBatchTest)
> {code:java}
> void testWithBatchSizeHavingSizeSameSizeAsStatements() {
>     def numRows = sql.rows("SELECT * FROM PERSON").size()
>     assert numRows == 3
>     def myOthers = ['f4':'l4','f5':'l5','f6':'l6','f7':'l7']
>     def result = sql.withBatch(myOthers.size(), "insert into PERSON (id, firstname, lastname)
values (?, ?, ?)") { ps ->
>         myOthers.eachWithIndex { k, v, index ->
>             def id = index + numRows + 1
>             ps.addBatch(id, k, v)
>         }
>     }
>     assert result == [1] * myOthers.size()
>     assert sql.rows("SELECT * FROM PERSON").size() == numRows + myOthers.size()
>     // end result the same as if no batching was in place but logging should show:
>     // FINE: Successfully executed batch with 4 command(s)
> }
> {code}



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)

Mime
View raw message