Problem with Textfield autocomplete -- only uses the INSTR function

Question

Hi dev team,
I use the new APEX "textfield autocomplete" item type and got a really damn problem with it.
I have a table with an index on the same column which I want to search with the autocomplete field.
Index:
CREATE BITMAP INDEX TEST.GEO_DATA_IDX1 ON TEST.GEO_DATA (postcode)
NOLOGGING TABLESPACE TS_INDEX;Select which is automatically generated by APEX 4:
SELECT   a.*
  FROM   (  SELECT   DISTINCT postcode AS RV
              FROM   TEST.GEO_DATA
             WHERE   country_id = :P1_COUNTRY
          ORDER BY   1) a
WHERE   INSTR ("RV", :p$_search_string) > 0 AND ROWNUM <= :p$_max_rows;
-- plan
Plan
SELECT STATEMENT  ALL_ROWSCost: 4,308  Bytes: 45,31  Cardinality: 1,97            
     2 SORT UNIQUE  Cost: 4,307  Bytes: 45,31  Cardinality: 1,97       
          1 TABLE ACCESS FULL TABLE #TABLE# Cost: 4,306  Bytes: 47,311  Cardinality: 2,057  In APEX 3 I used the addon from Tyler Muth with my own Select:
SELECT   DISTINCT
            postcode AS DV,
            postcode AS RV
     FROM   TEST.GEO_DATA
     WHERE   country_id = L_COUNTRY
     AND    postcode like l_search||'%'
     ORDER BY 1
-- Plan
SELECT STATEMENT  ALL_ROWSCost: 840  Bytes: 152,867  Cardinality: 13,897                                
     8 SORT ORDER BY  Cost: 840  Bytes: 152,867  Cardinality: 13,897                           
          7 HASH UNIQUE  Cost: 838  Bytes: 152,867  Cardinality: 13,897                      
               6 VIEW VIEW index$_join$_001 Cost: 836  Bytes: 232,111  Cardinality: 21,101                 
                    5 HASH JOIN            
                         2 BITMAP CONVERSION TO ROWIDS  Cost: 31  Bytes: 232,111  Cardinality: 21,101       
                              1 BITMAP INDEX RANGE SCAN INDEX (BITMAP) #INDEX#
                         4 BITMAP CONVERSION TO ROWIDS  Cost: 45  Bytes: 232,111  Cardinality: 21,101       
                              3 BITMAP INDEX SINGLE VALUE INDEX (BITMAP) #INDEX#Is there a chance getting my INDEX used by the new APEX item? I guess not. But I still want to ask before I rebuild everything. :)
Best regards,
Tobias

Answer 1

Hi Tobias,
I assume you are currently using "Contains & Case Sensitive" for the "Search" attribute. Have you already tried out "Exact & Case Sensitive"? Because that will internally use the LIKE operator as your original query does and because the % is at the end it should enable the use of an index. But it might still not work, because the ORDER BY might be executed at a different time than in your original example, but it's worth a try.
Regards
Patrick
My Blog: http://www.inside-oracle-apex.com
APEX 4.0 Plug-Ins: http://apex.oracle.com/plugins
Twitter: http://www.twitter.com/patrickwolf