Match Condition

Question

Hi,
I have two tables where I will be comparing a column that exists in both tables and that contains the same values but can be worded different e.g. (Case Sensitive, LTD Missing etc.)
I am using database version 11.2.0.3.0
Below is some sample data:
CREATE TABLE t1(
NAME VARCHAR2(255));
CREATE TABLE t2(
NAME VARCHAR2(255));
INSERT INTO T1 VALUES('ABC');
INSERT INTO T2 VALUES('ABC Ltd');
INSERT INTO T1 VALUES('Solutions UK Ltd');
INSERT INTO T2 VALUES('SOLUTIONS LTD');
INSERT INTO T1 VALUES('Associates Dawner Ltd');
INSERT INTO T2 VALUES('Dawner LTD');
INSERT INTO T1 VALUES('Equity (PVL)');
INSERT INTO T2 VALUES('EQUILIBRIUM PVL');
INSERT INTO T1 VALUES('Abacus PLC');
INSERT INTO T2 VALUES('Abacus T/A ABC');
INSERT INTO T1 VALUES('Asset Solutions Ltd');
INSERT INTO T2 VALUES('Asset (Kingsley) Ltd');
INSERT INTO T1 VALUES('Hunt Windows Ltd');
INSERT INTO T2 VALUES('Joe Hunt Ltd');
Can anyone help with a query that provides the closest match on the columns in both tables?
Thanks

Answer 1

Hi
You can play with function SOUNDEX. It give good result (>75% match). Example:
select n1, n2
from (select q.*,
             min(tp) over (partition by n1) mtp1,
             min(tp) over (partition by n2) mtp2
      from (select t1.name n1, t2.name n2, 1 tp
            from (select name, soundex(name) snd
                  from t1) t1,
                 (select name, soundex(name) snd
                  from t2) t2
            where t1.snd = t2.snd
            union all
            select t1.name, t2.name, 2
            from (select name, soundex(name) snd
                  from t1) t1,
                 (select name, soundex(name) snd
                  from t2) t2
            where t1.snd != t2.snd
            and substr(t1.snd, 1, 1) = substr(t2.snd, 1, 1)
            and substr(t1.snd, 2) between substr(t2.snd, 2) - 10 and substr(t2.snd, 2) + 10
            union all
            select t1.name, t2.name, 3
            from (select name, soundex(name) snd
                  from t1) t1,
                 (select name, soundex(name) snd
                  from t2) t2
            where t1.snd != t2.snd
            and substr(t1.snd, 1, 1) != substr(t2.snd, 1, 1)
            and substr(t1.snd, 2) between substr(t2.snd, 2) - 10 and substr(t2.snd, 2) + 10) q) q
where q.mtp1 = q.mtp2
and q.mtp1 = q.tp
Abacus PLC               Abacus T/A ABC
ABC                       ABC Ltd
Asset Solutions Ltd       Asset (Kingsley) Ltd
Equity (PVL)               EQUILIBRIUM PVL
Hunt Windows Ltd       Joe Hunt Ltd
Solutions UK Ltd          SOLUTIONS LTDEdited by: ---Brodyaga--- on Mar 5, 2013 11:27 PM
Edited by: ---Brodyaga--- on Mar 6, 2013 12:34 AM