Need help in optimizing the process

Question

I need help in optimizing my process. Here is what am I doing:
I have two tables table A and table B. table A has list of names and table B has name and ID. Right now I am loading name into cursor from table A and for each name from the cursor I try to match with name in table B to get the ID. This process takes very long to complete. I need help to replace PL/SQL with SQL if possible. I am using oracle utl_match to match the best possible record in table B. table B sometimes returns multiple matching records. I take the top nearest match name and ID for my use (there is not always one to one match)
Any idea or help will be appreciated. Thanks

Answer 1

always provide create table and sample data instert statements when asking a question. also, provide your database version.
because you're using utl_match, you can't really use an index, there must be a cartesian join on the two tables so it may always be slow.
however, here's a sample SQL that you may be able to make use of:
with n as (select 'KATHERINE' person_name from dual
             union all
             select 'STEVE' from dual
             union all
             select 'BRUCE' from dual)
    ,nid as (select 'CATHERINE' person_name, 1 NID FROM DUAL
             union all
             select 'STEFAN', 2 from dual
             UNION ALL
             select 'STEVEN', 2 from dual
             union all
             select 'CATHY',3 from dual)
select n_name, nid
  from (
        select n.person_name N_NAME
              ,nid.person_name
              ,nid.nid
              ,ROW_NUMBER() OVER (partition by n.person_name ORDER BY utl_match.jaro_winkler_similarity(n.person_name,nid.person_name) DESC) jws_ORDER
          from n
          join nid on (utl_match.jaro_winkler_similarity(n.person_name,nid.person_name) > 70)
  where jws_order = 1;