You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@jackrabbit.apache.org by "Benjamin Papez (JIRA)" <ji...@apache.org> on 2010/08/10 16:16:16 UTC

[jira] Created: (JCR-2704) Wrong SQL-2 parsing

Wrong SQL-2 parsing
-------------------

                 Key: JCR-2704
                 URL: https://issues.apache.org/jira/browse/JCR-2704
             Project: Jackrabbit Content Repository
          Issue Type: Bug
          Components: jackrabbit-jcr-commons, query, sql
    Affects Versions: 2.1.0, 2.0.0
            Reporter: Benjamin Papez


With the QOM factory I created the following constraints:

And
├ ChildNode
└ Or
     ├ Not
     │   └ PropertyExistence
     └ Comparison 

The created statement in the Query implementation object looks like this:

SELECT news.* FROM [jnt:news] AS news WHERE ISCHILDNODE(news, [/sites/jcrQueryTest/home/page8/news]) AND (NOT news.[jcr:language] IS NOT NULL OR news.[jcr:language] = 'en') ORDER BY news.[jcr:created] DESC

Here it would already be better if the QOMFormatter would translate Not -> PropertyExistence to "news.[jcr:language] IS NULL" instead of "NOT news.[jcr:language] IS NOT NULL" .

Anyway, for testing I took the above query statement and called:

                Query q = queryManager.createQuery(queryStatement, Query.JCR_SQL2)

and then I looked at the created statement and it converts to:

SELECT news.* FROM [jnt:news] AS news WHERE ISCHILDNODE(news, [/sites/jcrQueryTest/home/page8/news]) AND NOT (news.[jcr:language] IS NOT NULL OR news.[jcr:language] = 'en') ORDER BY news.[jcr:created] DESC

...AND (NOT news.[jcr:language] IS NOT NULL OR news.[jcr:language] = 'en')...
got converted to
...AND NOT (news.[jcr:language] IS NOT NULL OR news.[jcr:language] = 'en') ...

This looks like a bug in the Parser.

Furthermore if I take the following query:

SELECT news.* FROM [jnt:news] AS news WHERE ISDESCENDANTNODE(news, [/sites/jcrQueryTest]) AND (LENGTH(news.[j:nodename]) > 100 OR news.[date] > '+2001-01-01T01:02:03.000Z') AND (NOT news.[jcr:language] IS NOT NULL OR news.[jcr:language] = 'en')

and after queryManager.createQuery(queryStatement, Query.JCR_SQL2) the statement will be converted to:

SELECT news.* FROM [jnt:news] AS news WHERE ISDESCENDANTNODE(news, [/sites/jcrQueryTest]) AND (LENGTH(news.[j:nodename]) > 100 OR news.date > '+2001-01-01T01:02:03.000Z') AND (NOT news.[jcr:language] IS NOT NULL OR news.[jcr:language] = 'en')

So, suddenly the brackets around date are removed:

...OR news.[date] > '+2001-...
becomes
OR news.date > '+2001-

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

[jira] Resolved: (JCR-2704) Wrong SQL-2 parsing

Posted by "Thomas Mueller (JIRA)" <ji...@apache.org>.
     [ https://issues.apache.org/jira/browse/JCR-2704?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Thomas Mueller resolved JCR-2704.
---------------------------------

    Resolution: Duplicate

It looks like a duplicate of JCR-2564

> Wrong SQL-2 parsing
> -------------------
>
>                 Key: JCR-2704
>                 URL: https://issues.apache.org/jira/browse/JCR-2704
>             Project: Jackrabbit Content Repository
>          Issue Type: Bug
>          Components: jackrabbit-jcr-commons, query, sql
>    Affects Versions: 2.0.0, 2.1.0
>            Reporter: Benjamin Papez
>
> With the QOM factory I created the following constraints:
> And
> ├ ChildNode
> └ Or
>      ├ Not
>      │   └ PropertyExistence
>      └ Comparison 
> The created statement in the Query implementation object looks like this:
> SELECT news.* FROM [jnt:news] AS news WHERE ISCHILDNODE(news, [/sites/jcrQueryTest/home/page8/news]) AND (NOT news.[jcr:language] IS NOT NULL OR news.[jcr:language] = 'en') ORDER BY news.[jcr:created] DESC
> Here it would already be better if the QOMFormatter would translate Not -> PropertyExistence to "news.[jcr:language] IS NULL" instead of "NOT news.[jcr:language] IS NOT NULL" .
> Anyway, for testing I took the above query statement and called:
>                 Query q = queryManager.createQuery(queryStatement, Query.JCR_SQL2)
> and then I looked at the created statement and it converts to:
> SELECT news.* FROM [jnt:news] AS news WHERE ISCHILDNODE(news, [/sites/jcrQueryTest/home/page8/news]) AND NOT (news.[jcr:language] IS NOT NULL OR news.[jcr:language] = 'en') ORDER BY news.[jcr:created] DESC
> ...AND (NOT news.[jcr:language] IS NOT NULL OR news.[jcr:language] = 'en')...
> got converted to
> ...AND NOT (news.[jcr:language] IS NOT NULL OR news.[jcr:language] = 'en') ...
> This looks like a bug in the Parser.
> Furthermore if I take the following query:
> SELECT news.* FROM [jnt:news] AS news WHERE ISDESCENDANTNODE(news, [/sites/jcrQueryTest]) AND (LENGTH(news.[j:nodename]) > 100 OR news.[date] > '+2001-01-01T01:02:03.000Z') AND (NOT news.[jcr:language] IS NOT NULL OR news.[jcr:language] = 'en')
> and after queryManager.createQuery(queryStatement, Query.JCR_SQL2) the statement will be converted to:
> SELECT news.* FROM [jnt:news] AS news WHERE ISDESCENDANTNODE(news, [/sites/jcrQueryTest]) AND (LENGTH(news.[j:nodename]) > 100 OR news.date > '+2001-01-01T01:02:03.000Z') AND (NOT news.[jcr:language] IS NOT NULL OR news.[jcr:language] = 'en')
> So, suddenly the brackets around date are removed:
> ...OR news.[date] > '+2001-...
> becomes
> OR news.date > '+2001-

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