rigid seal
Is there a more efficient backfill than this approach?
(backfilling info directly throws the function read limit error)
backfillForAllUsers:
users = all users
for user in users:
schedule backfillForOneUser
backfillForOneUser:
infos = this user's info[]
for info in infos:
patch info
dense musk
Does efficient mean faster here, or are you trying to cut down on use of some other resource?
dense musk
This looks pretty reasonable, if you wanted to use fewer function calls or try to push its performance you could do some batching
but this looks like exactly how I'd write it, and how I hope developers write their code on Convex: it expresses what needs to be done and lets Convex handle the scaling
lusty coyote
In my migration helpers, I run each mutation in batches of 100 but expose a "change one document" function to make the API easy. https://stack.convex.dev/migrating-data-with-mutations
