db-derby-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Mamta A. Satoor (JIRA)" <j...@apache.org>
Subject [jira] Updated: (DERBY-3926) Incorrect ORDER BY caused by index
Date Wed, 20 May 2009 06:18:45 GMT

     [ https://issues.apache.org/jira/browse/DERBY-3926?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Mamta A. Satoor updated DERBY-3926:
-----------------------------------

    Attachment: DERBY3926_patch4_051519_stat.txt
                DERBY3926_patch4_051519_diff.txt

Army, thanks for looking at the patch. I need to remove the commented out if condition along
with corresponding {} brackets. Also, I have added the comment for currentRowOrdering.alwaysOrdered(previousOptimizable.getTableNumber())
Let me know if I should reword it differently. 

I have attached a patch with these changes and some more order by tests. That is the only
difference between this patch and the earlier patch.,

On another note, I was planning on moving all this new code inside the existing method OrderByList.sortRequired
which gets called by OptimizerImpl. This way, all the decision regarding sort avoidance would
have been made in the appropriate named method sortRequired. But the new code added by me
needs access to current join order position, previous optimizables, predicate lists etc which
are not available to OrderByList.sortRequired method. Passing these additional parameters
to the method will require changes to other part of the Derby code where we call this method.
Based on that, I am thinking that I should leave my new code where it is right now. If anyone
feels differently about the location of the new code, do let me know.

The only issue with the suggested patch is that a query like following is now going to require
a sort node on the top which we didn't require earlier. Other than this one case, all the
other test cases
have worked fine with my patch. I am inclied on going ahead and committing the patch with
this know one case. If I don't hear anything back on the patch, I will plan on comitting it
towards the end of the week.
select * from --DERBY-PROPERTIES joinOrder=FIXED 
TENKTUP1 -- DERBY-PROPERTIES index=TK1UNIQUE1 
, TENKTUP2 -- DERBY-PROPERTIES index=TK2UNIQUE1 
where TENKTUP1.unique1 = TENKTUP2.unique1 
order by TENKTUP1.unique1, TENKTUP2.unique1; 

> Incorrect ORDER BY caused by index
> ----------------------------------
>
>                 Key: DERBY-3926
>                 URL: https://issues.apache.org/jira/browse/DERBY-3926
>             Project: Derby
>          Issue Type: Bug
>          Components: SQL
>    Affects Versions: 10.1.3.3, 10.2.3.0, 10.3.3.1, 10.4.2.0
>            Reporter: Tars Joris
>            Assignee: Mamta A. Satoor
>         Attachments: d3926_repro.sql, derby-reproduce.zip, DERBY3926_notforcheckin_patch1_051109_diff.txt,
DERBY3926_notforcheckin_patch1_051109_stat.txt, DERBY3926_notforcheckin_patch2_051109_diff.txt,
DERBY3926_patch3_051509_diff.txt, DERBY3926_patch3_051509_stat.txt, DERBY3926_patch4_051519_diff.txt,
DERBY3926_patch4_051519_stat.txt, script3.sql, script3WithUserFriendlyIndexNames.sql, test-script.zip
>
>
> I think I found a bug in Derby that is triggered by an index on a large column: VARCHAR(1024).
I know it  is generally not a good idea to have an index on such a large column.
> I have a table (table2) with a column "value", my query orders on this column but the
result is not sorted. It is sorted if I remove the index on that column.
> The output of the attached script is as follows (results should be ordered on the middle
column):
> ID                  |VALUE        |VALUE
> ----------------------------------------------
> 2147483653          |000002       |21857
> 2147483654          |000003       |21857
> 4294967297          |000001       |21857
> While I would expect:
> ID                  |VALUE        |VALUE
> ----------------------------------------------
> 4294967297          |000001       |21857
> 2147483653          |000002       |21857
> 2147483654          |000003       |21857
> This is the definition:
> CREATE TABLE table1 (id BIGINT NOT NULL, PRIMARY KEY(id));
> CREATE INDEX key1 ON table1(id);
> CREATE TABLE table2 (id BIGINT NOT NULL, name VARCHAR(40) NOT NULL, value VARCHAR(1024),
PRIMARY KEY(id, name));
> CREATE UNIQUE INDEX key2 ON table2(id, name);
> CREATE INDEX key3 ON table2(value);
> This is the query:
> SELECT table1.id, m0.value, m1.value
> FROM table1, table2 m0, table2 m1
> WHERE table1.id=m0.id
> AND m0.name='PageSequenceId'
> AND table1.id=m1.id
> AND m1.name='PostComponentId'
> AND m1.value='21857'
> ORDER BY m0.value;
> The bug can be reproduced by just executing the attached script with the ij-tool.
> Note that the result of the query becomes correct when enough data is changed. This prevented
me from creating a smaller example.
> See the attached file "derby-reproduce.zip" for sysinfo, derby.log and script.sql.
> Michael Segel pointed out:
> "It looks like its hitting the index ordering on id,name from table 2 and is ignoring
the order by clause."

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


Mime
View raw message