18

While trying to get a handle on PendingEditRepBonus in the Data Explorer, I stumbled across an undocumented mystery in the DB. For example, thousands of posts with no "valid" PostHistoryType.

According to the "Database schema documentation for the public data dump and Data Explorer", and this answer to "Magic number in Data Explorer should be documented", there are only 22 PostHistoryTypes.

Yet the current DE snapshot shows:

PostHistoryTypeId    Name                    Count
-----------------    --------------------   -------
       1             Initial Title          1846229
       2             Initial Body           6016252
       3             Initial Tags           1846229
       4             Edit Title              223205
       5             Edit Body              2070256
       6             Edit Tags               430688
       7             Rollback Title            2612
       8             Rollback Body            10558
       9             Rollback Tags             4323
      10             Post Closed              36189
      11             Post Reopened             1752
      12             Post Deleted             31862
      13             Post Undeleted           31842
      14             Post Locked               2728
      15             Post Unlocked              211
      16             Community Owned         105436
      17             Post Migrated             4201
      18             Question Merged           3661
      19             Question Protected        1874
      20             Question Unprotected        58
      21             Post Dissociated            57
      23             null                        29
      24             null                     63936
      26             null                        13
      27             null                        80
      28             null                        57
      29             null                        37
      30             null                         1

What are these mysterious values for PostHistoryType?

Improve this question

1 Answer 1

Reset to default
12

Peeking around at the comments and posts associated with the revisions that have those types, I've made the following (not particularly insightful) guesses:

PostHistoryTypeId    Possible Purpose
-----------------    --------------------
       23            Unknown dev related event
       24            Approved suggested edit revision
       26            Vote nullification by dev (erm?)
       27            Post unmigrated/hidden moderator migration?
       28            Unknown suggestion event
       29            Unknown moderator event (possibly de-wikification?)
       30            Unknown event (too rare to guess)

With the possible exception of the suggested edit approvals, I feel like these revisions should have been excluded from the data. I could be wrong, though.

From Kevin Montrose:

Most of these weren't meant to be included (woops!), but aren't crazy important; just cleaning up bad data, generally. They probably won't be included in the next dump.

26 is a vote deletion record, used for things like "double upvote/accept/downvote got through" or "invalid/phantom bounty was started/awarded". It's pretty rare (being prompted by transient or one-off bugs), and doesn't leave much of a trace on the system. 10k+ user's can see one here, since the "close box" is not synced when votes are cleaned up.

Improve this answer
4
  • 1
    I suggest "Act of God" for 30 :)
    – Benjol
    Commented Aug 17, 2011 at 5:49
  • 3
    Most of these weren't meant to be included (woops!), but aren't crazy important; just cleaning up bad data, generally. They probably won't be included in the next dump. 26 is a vote deletion record, used for things like "double upvote/accept/downvote got through" or "invalid/phantom bounty was started/awarded". It's pretty rare (being prompted by transient or one-off bugs), and doesn't leave much of a trace on the system. 10k+ user's can see one here, since the "close box" is not synced when votes are cleaned up.
    – Kevin Montrose
    Commented Aug 17, 2011 at 6:15
  • 1
    @KevinMontrose Cool, thanks for the confirmation. Any hint on what 30 is for, for curiosity's sake? Or should we just go with Benjol's suggestion of "Act of God"? ;)
    – Tim Stone
    Commented Aug 17, 2011 at 6:44
  • 3
    @Kevin Montrose: Can we at least keep type 24? The DE data seems to have discrepancies compared to the actual rep audit. The biggest discrepancy is over edit-rep. and I'm hoping that history type might help narrow this down at some point.
    – Awesome Poodles
    Commented Aug 17, 2011 at 8:14

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .