Performance of JOIN vs. singleton SELECTs

Riley

DB2 v8 for z/OS
Data for each employee is spread across 12 tables.
We are accessing 12 tables with each table having from zero to one row per employee, except the first table has a required row for each employee.

What is the percentage of improvement if we do a JOIN of all 12 tables using
LEFT OUTER JOIN
vs. coding 12 singleton SELECTs ?

What if we have 5 tables in the above example ?

Aug 6 '07 #1

kalexin

There is no simple answer to this question. In theory there will be significant improvement if you have one statement as opposed to 12 statements, but there's no real way we can figure out a percentage without more information. Table statistics, parallelism on/off, etc etc.

Aug 9 '07 #2