This Week's Users includes users with no activity


(Bcguy) #1

I was looking through my list of “This Week”'s users - and saw a number of users on the list with nothing but zeros in all the columns. This seems like a bug - because if they were null on all the measures - it doesn’t seem like they should be on the list of this week’s users.

Am I missing something here? (See example below)

https://www.evernote.com/shard/s106/sh/acca5948-e79f-4a85-8211-2d762017b6ad/57e954fbbddf2e084b9d073e7dfa340c/deep/0/Users%20-%20Schizophrenia%20Forums.png


(Robin Ward) #2

Originally I used to exclude them but some argued that they couldn’t find people and wanted them in there even with no activity.


(Bcguy) #3

Well OK - I guess I’m unclear on why that would be, for “This week’s Users” list.

But - whatever the case - it seems that the zero column users should be grouped at the bottom and not sprinkled throughout the list rather randomly as currently seems to be the case.


(Tobias Eigen) #4

I was just about to report this as a bug but see you appear to have done this intentionally:

I think this is not correct behavior and breaks what could be a really useful feature to help members and moderators see who participated today or this week etc. This is valuable, actionable information. What’s the rationale here really and would you be willing to reconsider?


(Jeff Atwood) #5

Can you clarify something @eviltrout do you mean that users who at least visited the site are shown, even if they have not posted or liked or done anything that was countable?


(Robin Ward) #6

All user accounts are included. That means they could be included if the account was auto generated (API, import, etc) even if the user never logged in.


(Jeff Atwood) #7

Hmm that does seem very odd. Why include users who never visited the site? Was this an old customer requirement from a customer we don’t have any more?


(Robin Ward) #8

It’s been so long I don’t even remember. I just reviewed the query and the user has to be active which means a validated email - that would have to be faked quite explicitly from an import.

I guess the big question is: is a row of zeros desirable or should we exclude them?