Activations from same user counted multiple times in a cohort

[alexhanh]

CohortMe.analyze(period: 'months', activation_class: Transfer.unscoped) results in this query

SELECT DISTINCT transfers.user_id, transfers.created_at, cohort_date, FLOOR(extract(epoch from (transfers.created_at - cohort_date))/2592000) as periods_out FROM "transfers" JOIN (SELECT transfers.user_id, MIN(transfers.created_at) as cohort_date FROM "transfers" WHERE (created_at > '2014-04-18 11:31:53.062696') GROUP BY "transfers"."user_id") AS cohorts ON transfers.user_id = cohorts.user_id WHERE (created_at > '2014-04-18 11:31:53.062696')

If I make couple of transfers as a user in a cohort, then the count will increase by the number of transfers (activations) instead of just 1. That is, starting with this [2, 0, 0, 0, 0, 0, 0, 0, 0, 0] and then after multiple transfer by the same user results in [2, 0, 0, 0, 0, 0, 0, 0, 2, 0] instead of [2, 0, 0, 0, 0, 0, 0, 0, 1, 0].

I'm using Rails 4.2.1 and Postgresql 9.4.1. Please advice!

[procommerz]

Better late than never, for anyone who bumps into this (sorry for not posting a commit):

For Rails 4.2+ this line https://github.com/n8/cohort_me/blob/master/lib/cohort_me.rb#L54 has to to read like:

data.all.to_a.uniq {|d| [d.send(activity_user_id), d.cohort_date, d.periods_out] }

Otherwise ActiveRecord's relation uniq method takes precedence and the closure is not actually applied.