Re: Native DB command that populates tables

Mark

Ran a google on "automated test data generation" and got a bunch of stuff - some is pointed specifically at SQL Server, but those can be exported to iSeries if you have SQL Server sitting around (I find SQL Server 2000, using DTS, does it better than SQL Server 2005, using SSIS - the latter seems only to use OLE/DB providers, the former allows ODBC, and none of IBM's providers seemed to work for me on a customer project - also have seen several posts on the 'Net complaining about this - but did not have time to find out more).

On iSeries the only player I know of is Original Software - www.origsoft.com - they have an Extractor product that can give you subsets of your own data, tailored as you want, as I understand. All I know is what I read on their site.

I worked at a prominent DB vendor a few years ago that used a standard test suite for performance testing of its engine. They used TPC-H - we have heard of TPC-C worklloads and all - there is source code for a program (DBGEN) in all those TPC benchmarks that generates a set of tables (parts, suppliers, orders, etc.) with certain characteristics, and you specify how big a data set you want. The code is in C. It is freely available at http://www.tpc.org/tpch where you can see other benchmark sets, as well.

HTH
Vern

-------------- Original message --------------
From: "Mark Villa" <iseries.4.me@xxxxxxxxx>

Hi All,

I am curious if anyone else has checked deeper into native DB commands
from various vendors that could automatically populate new tables? And
once you establish relationships for the tables, the command observes
the schema rules.

The question: are there DB or vendor add-ons that you know of that
will generate pseudo data for load testing tables?

The ideal scenario might be running a DB command LOADSCH using a
starting key set, ending key set, and all possible combinations
between and perhaps starting at a 20% load rate. My idea of 20% is: DB
maintains good breadth of key ranges but saturation of possible column
contents only contains a random portion you targeted, i.e., 20%

I have not done the math, but I am guessing that a 60-100% load rate
would typically be millions of records.

This command is (in my view) "data processing" and it is what
computers do best and seemingly could be performed best by the native
DB manager.

The alternative seemingly is to write programs to load/reload or skimp
on test data volume and range quality.

--
Mark Villa
Summerville, SC
--
This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing list
To post a message email: MIDRANGE-L@xxxxxxxxxxxx
To subscribe, unsubscribe, or change list options,
visit: http://lists.midrange.com/mailman/listinfo/midrange-l
or email: MIDRANGE-L-request@xxxxxxxxxxxx
Before posting, please take a moment to review the archives
at http://archive.midrange.com/midrange-l.