Re: Creating test data

From: Jeremy Smith <alceste_at_xmission.xmission.com>
Date: Mon, 11 Nov 2002 21:53:40 +0000 (UTC)
Message-ID: <aqp8t4$cbq$1_at_terabinaries.xmission.com>


>> Is there anyone that knows about a program that can create test data for
>> database tables?

As far as I know, there isn't a full featured product. There's this website: http://www.realisticrandomdata.com/ which probably could be used in some cases.

However, if you're going to be doing performance or scalability testings, you're going to want to have data that looks like what you're really going to have in your system.

If you can't get a good amount of real data, one reliable method is to get some sample data, and if you need to scale it up, write simple programs that munges the data you're given in a fairly random way. And reinsert it. However, your fake data isn't going to have exactly the same properties.

I believe a company could make a good business out of providing a tool for testing databases that was able to take statistical properties of a database and create semi-random data with the same properties. I've thought about how to do that, it would actually require a good deal of effort. For instance, if I had a customer contact database, there would be trends in its data that made big performance differences, depending on whether I was mainly regional national, or global company.

If anyone has heard of such a product I'd be interested in it.

Jer Received on Mon Nov 11 2002 - 22:53:40 CET

Original text of this message