Log file Viewer

[08:08:57] <Joaolvcm[m]> hi, i have a list of rows unique by multiple fields, for now i was using updateOne+upser with a filter on those fields and hoping for the best, but that sometimes create duplicate fields. Should i create a custom _id composed by the fields i'm filtering on, or should i use unique compound index? the thing is that in this case duplication is better than error/data loss, and i'm unsure if unique compound index will throw

[08:08:57] <Joaolvcm[m]> an error and if there is any disadvantage or using custom _id

[16:10:14] <hardwire> are you wrapping that in a transaction?

[16:10:19] <hardwire> That could be the ticket.