Generating Test Data For Your Application

devcurry.com

13 points·brianclintwud··4 comments

The feature of being able to describe test data sets, and then generate what really comes down to SQL is really neat. And it looks like they offer SQL exporting.

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.

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.

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

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...

Reply on news.ycombinator.com