From: Hoo man <hoo@online.de>

Date: Mon, 02 Jul 2012 19:25:43

Ok, I got a few questions regarding this ticket:
First of all, is it ok, if I exclude users which are blocked? There might be privacy issues with the names etc. Same goes for deleted revisions in 3.
Then for the third, is it ok, if it only includes the users from 1? As your query doesn't do this atm.

Finally, no I can't dump simplewiki_p (alone the revision database will have several GiB, I guess), I'll just write the queries myself, as they are rather trivial and that should be faster for me than dumping a wiki.

From: Gregor Martynus <gregor@martynus.net>

Date: Mon, 02 Jul 2012 19:56:44

thanks a lot for your quick response!

First of all, is it ok, if I exclude users which are blocked? There might be privacy issues with the names etc. Same goes for deleted revisions in 3.

Yes, sure, you can exclude users that are blocked and revisions that have been deleted.

Then for the third, is it ok, if it only includes the users from 1? As your query doesn't do this atm.

Absolutely, actually I'd prefer that, I just didn't want to make the query to complex.

Thanks a lot for your help!

From: Hoo man <hoo@online.de>

Date: Mon, 02 Jul 2012 23:06:43

As told, this was an easy one, so here we go:
SQL:

--1
SELECT user_id, user_name, user_registration FROM user INNER JOIN logging ON log_user = user_id WHERE LEFT(user_registration, 4) = 2012 AND user_id NOT IN (SELECT ipb_user FROM ipblocks) AND log_type = 'newusers' AND log_action = 'create';
--2
SELECT page_id, page_title, page_namespace, page_is_redirect FROM page;
--3
INSERT INTO u_hoo.dbq189 SELECT user_name FROM user INNER JOIN logging ON log_user = user_id WHERE LEFT(user_registration, 4) = 2012 AND user_id NOT IN (SELECT ipb_user FROM ipblocks) AND log_type = 'newusers' AND log_action = 'create';
SELECT rev_id, rev_page, rev_comment, rev_user_text, rev_user, rev_timestamp FROM revision INNER JOIN u_hoo.dbq189 ON rev_user_text = dbq189.user_name WHERE rev_deleted = 0 AND rev_user != 0;

Result:
http://toolserver.org/~hoo/dbq/dbq-189_0.txt (plain text)
http://toolserver.org/~hoo/dbq/dbq-189_1.txt (plain text)
http://toolserver.org/~hoo/dbq/dbq-189_2.txt (plain text)

From: Gregor Martynus <gregor@martynus.net>

Date: Tue, 03 Jul 2012 23:15:07

thanks Hoo man, you saved me so much work, really appreciate your help!!

From: Gregor Martynus <gregor@martynus.net>

Date: Wed, 04 Jul 2012 16:57:05

sorry to bother you once more,
but there seem data to be missing.

Example:

SELECT * FROM revisions WHERE rev_user = 1426164;

Returns 5 results. But when I open http://de.wikipedia.org/wiki/Spezial:Beitr%C3%A4ge/Johannes_Michels I see way more contributions by this user. Does that make any sense to you?

From: Hoo man <hoo@online.de>

Date: Wed, 04 Jul 2012 19:42:00

hoo@yarrow:~/public_html/dbq$ cat dbq-189_2.txt | grep -P '\t1426164\t' | wc -l

32

So the problem seems to be on your side, maybe you're not importing the data properly?

From: Gregor Martynus <gregor@martynus.net>

Date: Thu, 05 Jul 2012 00:02:40

you are right, sorry. I used a MAC app, I didn't get any errors, but data was indeed missing. Now I used mysqlimport and it works. Thanks again!

From: Gregor Martynus <gregor@martynus.net>

Date: Fri, 06 Jul 2012 00:00:44

Quick question regarding reverted revisions:

I asked initially for the rev_deleted column because I thought it would be the flag for reverted revisions, but I guess that's not the case, right?

Is there any chance to get an additional flag if a revision has been reverted? If yes, could you add it to the dbq-189_2.txt dump? It would really help a lot for my friend's study

From: Hoo man <hoo@online.de>

Date: Fri, 06 Jul 2012 00:04:14

Sadly, that's not possible, rev_deleted indeed stores whether the revision has been deleted and not if it has been reverted.

From: Gregor Martynus <gregor@martynus.net>

Date: Fri, 06 Jul 2012 00:07:10

Thanks for the quick response!

I'm just wondering: how can a revision be highlighted as reverted, when there is now such flag in the database? Is there a trick I could use myself to identify revisions that have been reverted?

From: Hoo man <hoo@online.de>

Date: Fri, 06 Jul 2012 00:11:25

The only way I see at the moment would be to get follow-up revisions and check whether they're reverts, getting those shouldn't be to complicated, but detecting reverts will be more hard. And the data gathered that way wont be fully reliable either, as there are various cases in those you wont be able to detect reverts that way.

From: Gregor Martynus <gregor@martynus.net>

Date: Fri, 06 Jul 2012 00:13:55

okay, thanks a lot for your instant help!

From: Gregor Martynus <gregor@martynus.net>

Date: Thu, 06 Sep 2012 23:22:31

May I ask you to run the same 3 queries from the comment above and provide me the results? Thanks for your help!

This bug was imported as RESOLVED. The original assignee has therefore not been
set, and the original reporters/responders have not been added as CC, to
prevent bugspam.

If you re-open this bug, please consider adding these people to the CC list:
Original assignee: hoo@online.de
CC list: hoo@online.de