I am hoping we can recover the garbled usernames from the raw JSON logs, but you’re correct about username changes. For project level counts, though, they should not dramatically affect the accuracy of new registration numbers.

On Jun 5, 2014, at 3:51 PM, Aaron Halfaker <ahalfaker@wikimedia.org> wrote:

Regretfully, looking up a user in Centralauth requires the use of a username.   Then again, you'd need to join with a user table (with user_id) anyway since users can be renamed after they create their account and that name change won't be reflected in ServerSideAccountCreation.  


On Thu, Jun 5, 2014 at 5:47 PM, Steven Walling <swalling@wikimedia.org> wrote:

On Thu, Jun 5, 2014 at 1:24 PM, Dario Taraborelli <dtaraborelli@wikimedia.org> wrote:

• Use event_userId whenever possible

This is really a best practice everyone should follow in all analysis. Unless you're qualitatively interested in the contents of usernames, any analysis that uses unique names instead of ids should probably be treated as highly suspect. 


--
Steven Walling,
Product Manager

_______________________________________________
Analytics mailing list
Analytics@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/analytics


_______________________________________________
Analytics mailing list
Analytics@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/analytics