# Testing with fixed data set

**URL:** <https://discourse.getdbt.com/t/testing-with-fixed-data-set/564>\
**Category:** In-Depth Discussions\
**Tags:** testing\
**Created:** [September 10, 2019, 5:47am UTC](https://discourse.getdbt.com/t/testing-with-fixed-data-set/564 "2019-09-10T05:47:33Z")\
**Posts on this page:** 1\
**Showing post:** 8

<div class="post-metadata">

**Author:** ![jmriego](https://avatars.discourse-cdn.com/v4/letter/j/d9b06d/32.png) [@jmriego](https://discourse.getdbt.com/u/jmriego)\
**Post date:** [March 4, 2021, 1:51pm UTC](https://discourse.getdbt.com/t/testing-with-fixed-data-set/564/8 "2021-03-04T13:51:02Z")

</div>

Definitely agree with you on that @clee , I both think unit testing is really important for mature use of the project and tests should not requrie modifying the existing DBT models.

The example I put there doesn’t really use a super-ref macro but it still requires modifying the existing ref function. Using your example of testing Model C which has depends on Source A and Model B, the testing scripts will run dbt jobs but passing it variables looking like this: `{'A': 'testcase1', 'B': testcase1'}`

So the ref function doesn’t guess where the test case should be but you can configure it at run time. As you I didn’t find a perfect solution that would allow fakes but I think this is good enough for testing transformations

---

_[View the full topic](https://discourse.getdbt.com/t/testing-with-fixed-data-set/564)._