Many customer wonder why MettleCI shouldn’t just be used to capture The unit test data files are stored ion disk as CSV files. Footprint.
Those files need to be committed to Git to be used for unit testing as part of a CI build process. Footprint.
MettleCI output comparison is not order dependent, meaning …
the order of your actual output could be different to your expected output SV file, and
the order of your actual output could change between job
… and your test would still pass as long as every output row was present and identical to the rows a your expected CSV file.
Why doesn’t MettleCI allow us to minimize disk by just take the first n rows from each input?
It would be from full-volume executions as a quick means of verify your jobs' consistent behaviour. There are a number of reasons why this is not a valid of of MettleCI’s Unit Test capability. Please see Managing your Unit Test data volumes for a discussion of the reasons.