Last modified: 2014-03-11 13:32:38 UTC
On largest dumps this routine can take hours, not trivial to rework though
https://mingle.corp.wikimedia.org/projects/analytics/cards/344
One of the design anomalies that dates back to an era when the English dump could be parsed in minutes rather than days ;-) Current implementation is really inefficient (bad coded) as for some metrics in WikiCountsOutput.pm data are established month by month, and for every iteration a recursive lookup occurs. The inefficiency grows over month by month, but speed improvements in hardware make it less urgent. If Wikistats would change to incremental updates https://bugzilla.wikimedia.org/show_bug.cgi?id=46198 the issue is mute.