OHDSI Home | Forums | Wiki | Github

DQD test failure explanation is non-intuitive

First, let me say I’m quite impressed by DQD! What a great harness to be able to run and quantify defects in the OMOP payload. Below is a FAIL, however, that has non-intuitive wording describing why it failed. “The number and percent of records that have a value in the PROVIDER_ID field in the PROCEDURE_OCCURRENCE table that does not exist in the PROVIDER table is (next column) 0%.” What am I missing? That seems like a PASS. TIA.

1 Like

Hi there! Did you check the number of failed records in the expanded view below the check row? The actual failure percentage may be rounded to 0 in the UI but still non-zero under the hood.

Thank you Katy! There are 8 rows violated. Again, what a great tool. I worked at SAS for 22 years primarily in testing/validating numeric output. We had millions of tests and 500 dedicated testers. I have great appreciation for the work it took to make this framework!

Very glad to hear it! All credit goes to the illustrious @clairblacketer for the development of this indispensable tool :smiley:

t