Closed philbooth closed 7 years ago
It turns out I had unjustified optimism about how well the COPY
query would handle mixed-length CSV rows. So to make this work I had to pre-process 2017-02-23
CSV in awk to make this work correctly. With that change, plus the usual dosage of trial and error to fix the obligatory SQL syntax errors, this is now running for real. Merging.
Fixes #36. Fixes #48. Replaces #54.
This is rebased against #53 because merge conflicts. You might want to review that one before this one, or both together is fine too.
mozilla-services/puppet-config#2488 was merged last night, which means the CSV files going forward will contain
locale
anduid
fields. This change updates the import scripts so that we populate new columns for these fields when they're present.We could add the new columns to the existing
flow_metadata
schema with a separate query but, given that #53 needs doing at the same time, I'm minded to drop all the data and re-import it. Two birds, one stone, zero bespoke queries.@rfk r?