Putting this up as a possible way to handle the broken data without losing full games by throwing a warning and doing a left join if we have mismatched game and survey data. Requires manually assigning a new participant id and invite id to one of the duplicate participant records, which serves as a marker for where this occurred in the end
Also fixes the issues with survey processing (removing extra surveys for participants, switching to bind_rows() to revert to previous functionality)
Gets us out of this hole and I figure any further effort is probably better placed towards redoing this with pandas.. Lmk if this seems alright @alee, or if you have a diff solution already
Putting this up as a possible way to handle the broken data without losing full games by throwing a warning and doing a left join if we have mismatched game and survey data. Requires manually assigning a new participant id and invite id to one of the duplicate participant records, which serves as a marker for where this occurred in the end
Also fixes the issues with survey processing (removing extra surveys for participants, switching to
bind_rows()
to revert to previous functionality)Gets us out of this hole and I figure any further effort is probably better placed towards redoing this with pandas.. Lmk if this seems alright @alee, or if you have a diff solution already