topic Faster way to write a SQL Update Query in SAS Programming

Faster way to write a SQL Update Query

Reeza — Mon, 18 Aug 2014 21:24:11 GMT

Is there a more efficient way to write this SQL Update Query?

It has to be an update, not a table overwrite, though the update can be through SQL or a data step.

Thanks!

PROC SQL;

UPDATE TABLE1

SET VAR1 = (select P.VAR1 from TABLE2 as P where P.GROUP = I.GROUP and P.NUM = &ID)

, VAR2 = (select P.VAR2 from TABLE2 as P where P.GROUP = I.GROUP and P.NUM = &ID)

, VAR3 = (select P.VAR3 from TABLE2 as P where P.GROUP = I.GROUP and P.NUM = &ID)

, VAR4 = (select P.VAR4 from TABLE2 as P where P.GROUP = I.GROUP and P.NUM = &ID)

, VAR5 = (select P.VAR5 from TABLE2 as P where P.GROUP = I.GROUP and P.NUM = &ID)

;

QUIT;

Re: Faster way to write a SQL Update Query

Patrick — Mon, 18 Aug 2014 23:01:26 GMT

You could use the SAS UPDATE statement SAS(R) 9.4 Statements: Reference, Third Edition

You could load "table2" into a hash table, look-up the values and if there is a match to the hash replace the record using a SAS MODIFY statement with REPLACE (this way no sorting of table1 required).

For your SQL: You could subset table2 (where NUM=&ID) before using it in your sub-selects

Re: Faster way to write a SQL Update Query

Tom — Tue, 19 Aug 2014 02:27:47 GMT

Look into using the MODIFY statement.

Look at Example 3 in this help document.

SAS(R) 9.3 Statements: Reference

Re: Faster way to write a SQL Update Query

LinusH — Tue, 19 Aug 2014 10:28:31 GMT

I agree with Tom, MODIFY is usually the fastest way to do updates, given that the transaction table is relatively small.

If you wish do keep the SQL, be sure to have proper indexes (on both master and transaction tables), it could speed things up.