The DDL sequencer should allow specifying the dialect type in DDL comments
--------------------------------------------------------------------------
Key: DNA-641
URL:
https://jira.jboss.org/jira/browse/DNA-641
Project: DNA
Issue Type: Feature Request
Components: Sequencers
Affects Versions: 0.7
Reporter: Randall Hauch
Fix For: 1.0
At the moment, the DdlParsers class looks for a leading "PARSER_ID =
<DialectName>" at the beginning of the DDL stream. If this is not found, then
the class attempts to automatically discover the dialect by performing a best fit.
Unfortunately, the "PARSER_ID = <DialectName>" is not valid DDL, and thus
can never be included in a real DDL file.
Rather than look for this kind of leading statement, the DdlParsers should look for the
first comment block and simply attempt to find the dialect names in the comment text. This
should be done in a case-insensitive way, and would ideally count the number of
occurrences of each dialect name (thus handling the case where an Oracle DDL file
contained both "oracle" and "postgres" in the comment; hopefully the
"oracle" term would appear more than others). This of course is very imprecise,
but it is perhaps only 'better than' the use of non-DDL statements.
It is also desirable that the DdlParsers class should be tolerant of incorrectly typing
the DDL stream. For example, lets say a DDL file contains a comment containing
"oracle" and no other dialect name, but this DDL file is actually for PostgreSQL
and merely contains the word "oracle" because it was converted from another
Oracle DDL file. In this case, if parsing with the PostgreSQL parser results in an error,
the DdlParsers should start over, behaving as though the dialect typing was incorrect, and
should perform a "best fit".
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
https://jira.jboss.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see:
http://www.atlassian.com/software/jira