Add VAT validation rule #5 [VS-16] #7365
Conversation
| @@ -245,7 +258,7 @@ task SchemaOnlyOneRowPerNullTranscript { | |||
| transcript_source is NULL AND | |||
| transcript is NULL | |||
| GROUP BY vid | |||
| HAVING num_rows = 1' > bq_variant_count.csv | |||
| HAVING num_rows > 1' > bq_variant_count.csv | |||
There was a problem hiding this comment.
nit: while you are here, there's a tiny typo below:
"# if the result of the query has any rows, that means there were vids will null transcripts and multiple"
I think will should be with?
|
converting to draft because |
|
ah---this is what I get for making an example inputs file when we really didn't need one and for choosing the full AoU 1k release (vat_kc_vat_1) as the default. I wanted to run the validations here because it is the largest dataset and it is the AoU data (not just Anvil data) BUT it has no values for gvs_all_ac or gvs_all_an yet because that step wasn't implemented by the time of creation. (Validation #9 was added by Lee recently) The table vat_jul18 does have those values as it was created just last week, but may get cleaned up...so this might be a good best practices question for what we run this on in the future if there is ever an automated version? |
Closes https://broadworkbench.atlassian.net/browse/VS-16 by adding SQL for