dear *,
i have a rather theoretical question.
short version:
given all the queries to be executed over a given table schema, is there
a "calculus" for finding all the indices that should be created (to
avoid linear table scans) ?
long version:
i have a given set of tables. i know all the queries i want to run
against those tables (i do not know the sizes of the tables). how do i
tell which indices i should create ? which rules does a DBA use to
determine the indices ? can i put those rules into a computer program ?
i am especially unsure about the column order for a multi-column index,
given the WHERE-clause to apply.
is there a rule set ? does anyone know papers on that subject ? is there
a proof that there cannot be such a rule set ?
thank you very much in advance,
Raimund
--
Raimund Jacob
\|/ ____ \|/
Pinuts Media+Science GmbH "@'/ ,. \`@"
Potsdamer Str. 96 /_| \__/ |_\
D-10785 Berlin \__U_/
http://www.pinuts.de/ voice: +49 30 590090313 fax: +49 30 590090390
_______________________________________________
sapdb.general mailing list
[EMAIL PROTECTED]
http://listserv.sap.com/mailman/listinfo/sapdb.general
