How long is it going to take?

Question

Tables:
x_cart - 50,000 rows
x_1 | x_2
a11 | b11
c12 | g23
y_concat
x_all - 11,000,000 rows
a11_c12_g45
c12_g23
c12
a11_g23_b11_c12We need to count where x_1 and x_2 from x_cart table is found within x_all in y_concat table. So in example above we should get:
x_1 | x_2 | Total_count
a11 | b11 | 1
c12 | g23 | 2I have the following query to do this:
CREATE TABLE x_counts AS
  SELECT a.x_1,
         a.x_2,
         COUNT(1) AS total_count
  FROM   (SELECT x_1,
                 x_2
          FROM   x_cart) a
          JOIN y_concat b
           ON Instr (b.x_all, a.x_1) != 0
              AND Instr (b.x_all, a.x_2) != 0
  GROUP  BY a.x_1,
            a.x_2; But it is taking too long.
CREATE TABLE STATEMENT  ALL_ROWSCost: 341,329,480  Bytes: 94,386,107,624  Cardinality: 1,194,760,856Any suggestions to improve the code? If not can you roughly estimate how long it is gonna run?
Thanks
Edited by: zx8754 on 07-Jun-2010 05:00

Answer 1

CREATE TABLE x_counts AS
SELECT a.x_1,
a.x_2,
COUNT(1) AS total_count
FROM (SELECT x_1,
x_2
FROM x_cart) a
JOIN y_concat b
ON a.x_1 like 'b.x_all%'
AND a.x_2 like 'b.x_all%'
GROUP BY a.x_1,
a.x_2;
and create index over b.x_all
as like may use index id % is given on one side
No logging also may improve performace...