[rt-devel] Keyword strangeness.

[Matthew Watson]

Heya all.

 I actually noticed this along time ago but never really got a chance to 
look into it.
my RT2 database was converted from rt1, and in the process I think it 
fudged the keywords a bit

I seem to have duplicate Keywords in the database, eg.

mysql> select name,count(*) from Keywords group by Name;
...
| ACCOUNTING QUERY       |     1254 |
| CALL CENTRE            |   108228 |
etc..

Now, I would expect perhaps a duplicate or two from sub-keywords, but 
100,000?

now, most of these keywords are infact disabled.

mysql> select count(*)
    -> from Keywords,ObjectKeywords
    -> where ObjectKeywords.keyword=Keywords.id
    -> and Keywords.Disabled=1;
+----------+
| count(*) |
+----------+
|   206689 |
+----------+
1 row in set (2.05 sec)

So i'm guessing the import script went a bit wacky and created a new 
keyword for every ticket that used that keyword, rather than reusing the 
existing one.

Now, two questions,

1, would having 200,000 or so Disabled keywords effect the speed of RT much?

and 2, if so, any suggestions for fixing this, i've read that directly 
manipulating the DB is a pretty big no no in RT, but I guess this is
likely to be an exception.

Regards,
Matthew Watson
Netspace Online Systems.