Backfill distribution table

[dianaclarke]

No description provided.

[ElenaHenderson]

@dianaclarke If we only select summary rows with run_ids for runs with name like commit: %, we can cut down loop iterations (for summary in summaries) by more than 50%:

select count(1) 
from summary

2288167

select count(1) 
from summary, machine
where summary.machine_id = machine.id and
machine.name in ('ursa-i9-9960x', 'ec2-t3-large-us-east-2', 'ec2-t3-xlarge-us-east-2', 'ursa-thinkcentre-m75q') 

1525400


select count(1) 
from summary, machine, run
where summary.machine_id = machine.id and
machine.name in ('ursa-i9-9960x', 'ec2-t3-large-us-east-2', 'ec2-t3-xlarge-us-east-2', 'ursa-thinkcentre-m75q') and
summary.run_id = run.id and
run.name like 'commit: %'

993577


select count(1) 
from summary, run
where summary.run_id = run.id and
run.name like 'commit: %'

1036257