Accent insensitive + Search for 2 non consecutive keywords

Question

Hi,
In APEX, I have a query the search into a list of TITLE (TIT), with or without accents. I cannot search for 2 non-consecutive word in my title and have the result.
Here an example:
SELECT
"ID","TIT"
from   "REGDOSSIERS" "RDO"
where
instr(TRANSLATE(UPPER("TIT"),'ÁÂÈÉÊÍÓÚÜ','AAEEEIOUU'),upper(nvl(:P1_REPORT_SEARCH,"TIT")) ) > 0
TIT
Last name of the resolution
King Henry the Sixth
The Gazette
The Comedy of Errors
La vie enchanté
Le bêtes et les oiseaux1- If I look for: Last Name_, I have this result:
TIT
Last name of the resolution2- If I look for: Last resolution*,
I have NO RESULTS
My Question: How to I rewrite my query to have be able to look for 2 keywords non consecutive in the field,
thanks
Roseline, Quebec, Canada
Edited by: user8772975 on 2009-09-04 10:47
Edited by: user8772975 on 2009-09-04 10:49
Edited by: user8772975 on 2009-09-04 10:50
Edited by: Roseline on Oct 1, 2009 7:27 PM

Answer 1

Hi,
If you know the order of the keywords (that is, if 'LAST' must come before 'RESOLUTION'):
WHERE  str  LIKE '%LAST%RESOLUTION%'where str is the string with accents standardized.
Now, since you're starting with a variable like 'LAST RESOLUTION', you need to convert it to '%LAST%RESOLUTION%', so say
WHERE  str  LIKE '%' || REPLACE ( :p1_report_search     -- TRANSLATE would work, too
               || '%'Watch out for the "mother is in chemotherapy" problem.
That is, if str = 'AMY''S ORDEAL WITH CHEMOTHERAPY", and :p1_report_search = 'MY MOTHER', you'll get a match, because str contains the 2-character sibstring 'MY' and the 6-character substring 'MOTHER' in that order.
If that's not what you want, then you can search on whiole words only, like this:
{code}
WHERE ' ' || REPLACE ( str
          , ' '     -- 1 space ...
          , ' '     -- ... becomes 2 spaces
     || ' ' LIKE '% ' || REPLACE ( :p1_report_search     -- TRANSLATE won't work, now
                    , ' % '     -- spaces before and after wild-card
               || ' %'
{code}
This assumes that all words are delimited by spaces.
If you want punctuation or other whtiespace (like tabs) to delimit words, that can be done, too.
Edited by: Frank Kulash on Sep 4, 2009 2:09 PM
Added whole-word example.