Generating Test Data For Your Application(devcurry.com)
devcurry.com
Generating Test Data For Your Application
http://www.devcurry.com/2009/12/generating-test-data-for-your.html
2 comments
Not to mention there are already some good libraries in popular languages that are capable of doing most of the hard work:
Perl - Data::Faker - http://search.cpan.org/~jasonk/Data-Faker-0.07/lib/Data/Fake...
Ruby Gem - Faker - http://faker.rubyforge.org/
Bonus round!
Ruby Gem - Populator - http://github.com/ryanb/populator
Faker generates "fake" values, while Populator provides some nice constructs for using collections/ranges of data for your fields.
Perl - Data::Faker - http://search.cpan.org/~jasonk/Data-Faker-0.07/lib/Data/Fake...
Ruby Gem - Faker - http://faker.rubyforge.org/
Bonus round!
Ruby Gem - Populator - http://github.com/ryanb/populator
Faker generates "fake" values, while Populator provides some nice constructs for using collections/ranges of data for your fields.
I wrote something like this for myself a while back, then stuck it up on GitHub. Its written in Java and generates all sorts of "user" related data. Addresses, emails, etc. Can also generate social-graph connections between the generated users.
DummyDataGenerator: http://github.com/mindcrime/dummydatagenerator
DummyDataGenerator: http://github.com/mindcrime/dummydatagenerator
Since my app is meta-data driven, it was easier for me to implement an "Add Record with random values" for my testing. Exposed it for public users and I have see it being used a lot.
In action (animated gif, 186 KB) : http://crm.ifreetools.com/static/images/blog/generating-rand...
In action (animated gif, 186 KB) : http://crm.ifreetools.com/static/images/blog/generating-rand...
However, if you are generating test data for an application you've created - this really only comes in hand if you are lazy. I only say that because to really harness the (limited) power, it requires a "donation" of $20. Not that supporting the cause is bad or any of that jazz - I get it. But if I'm giving someone 20 dollars for it, it better be more than stub data, and don't limit me to 5000 records at a time either.
I'm not trivializing this tool - but for something that could be hacked in IRB in 10 minutes there has to be a more compelling reason to do it this way.