mozilla / mozilla-reports

Repository for public analyses.
https://mozilla.report
5 stars 12 forks source link

Bug 1379622 - Handle Dataset's auto-sanitized where clause. #68

Closed mreid-moz closed 7 years ago

mreid-moz commented 7 years ago

Dataset was recently changed to automatically sanitize arguments to the where clause. This caused us to never match any files, since the actual prefix contains a - rather than an _. Use the prefix parameter of Dataset instead.

mreid-moz commented 7 years ago

Re: renaming the dataset - I don't think so. It is quite common for the dataset itself to have a hyphen on S3 - the problem only arises due to my hacky approach of treating the prefix as fake dimensions (IIRC this ETL code pre-dates the 'prefix' feature of the Dataset API).