Talk:Mix'n'match/Archive 2

Request to remove/clear catalog 5860

Catalog 5860 is for P1733 on Wikidata, but it has bad data for "official website" which keeps being added to items. It's not particularly useful as a catalog itself, as we typically have a means of importing Steam IDs automatically based on a number of other external IDs anyway. I'd like it removed or cleared to prevent further imports of bad official website data. Nicereddy (talk) 16:34, 29 December 2023 (UTC)

Deactive catalog 195

Hi, because the source dataset has changed its identifiers, catalog 195 is currently creating/adding incorrect matches to Wikidata. While (per the corresponding Wikidata property talk page) several Wikimedians are in touch with the data publishers about bringing back the old identifiers, it's probably best to (hopefully temporarily) disable this catalog. Can someone do this? As the catalog creator I can't find a way to do it myself unfortunately. Pinging @Magnus Manske @Jean-Frédéric @Lucas Werkmeister - thank you! Spinster (talk) 15:44, 30 September 2024 (UTC)

Sure thing, Done (in https://mix-n-match.toolforge.org/#/catalog_editor/195) --Jean-Fred (talk) 17:04, 30 September 2024 (UTC)

New property for catalogue 6490

Could someone with access please update catalog 6490 with the recently-created property Western Australian Biographical Index (P13035)? Thanks! Sam Wilson 00:45, 4 October 2024 (UTC)

Just had a look, looks like it’s been done already. Jean-Fred (talk) 10:00, 9 October 2024 (UTC)

@Jean-Frédéric: Oh great! Thanks to whoever did it. :-) Sam Wilson 08:41, 10 October 2024 (UTC)

Deactive catalogs 868, 2273 and 713

Hi! This datasets, all related to argentina, changed their identifiers. To avoid creating more incorrect matches, can we deactivate or delete them?

• https://mix-n-match.toolforge.org/#/catalog/868 autores.ar died years ago.
• https://mix-n-match.toolforge.org/#/catalog/2273 Directorio Legislativo changed their system, identifiers, everything...
• https://mix-n-match.toolforge.org/#/catalog/713 this never really had PID, they were just blog post that someone misidentified and created a property...only lasted a year.

Thanks! --Mauricio V. Genta (talk) 15:14, 17 October 2024 (UTC)

Disabled all three. Thanks for reporting! Jean-Fred (talk) 18:24, 17 October 2024 (UTC)

That said, autores.ar ID (P4158) could have its formatter URL changed to the wayback machine, and then the catalogue updated accordingly. That was what was done for Argentinian Historic Heritage ID (P4587) already. Jean-Fred (talk) 18:26, 17 October 2024 (UTC)

Add property P13104 to catalog 6380

Property graphclasses.org ID (P13104) has been recently created. It should be added to 6380, which currently has no Wikidata property. Dexxor (talk) 10:59, 30 October 2024 (UTC)

Mix'n'match updates for Encyclopedia of Korean Culture catalog

Mix'n'match not updating Encyclopedia of Korean Culture correctly. The total count seems to be updating correctly but updates are incorrectly ascribed to "Auxiliary Data Matcher" rather than to (for example) "MargaretRDonald". Lately perhaps 1 in 10 are correctly attributed. MargaretRDonald (talk) 03:10, 30 November 2024 (UTC)

Please use new multiple language default for persons

Instead of setting pt-br fi sv gl nl pt it es de en fr only set mul for the label. See d:Help:Default values for labels and aliases. Matthias (talk) 19:27, 4 January 2025 (UTC)

Deactivate catalog 506, BHL Creator IDs, new one at 6686

Hi! People here at BHL have noticed that 506 still had many ids that are currently deprecated on the BHL end.

I've uploaded this new catalog at 6686 without the deprecated IDs. Is there any way to deactivate the old one so to avoid confusion? Cheers!

TiagoLubiana (talk) 18:22, 27 January 2025 (UTC)

@TiagoLubiana: deactivated 506; this page is (probably) not monitored, if you add a ping to me I see the request and proceed quickly. Having a dedicated page for these requests would be useful (most are send directly in d:User talk:Magnus Manske presently), cf. my message at d:Wikidata talk:Mix'n'match#Procedure and criteria for requests of catalogues' deactivation if you want to comment. Epìdosis 11:15, 6 April 2025 (UTC)

Delete catalog 6776

Please delete catalog 6776 -- I messed up the import, and it's currently useless. :) Thanks, and apologies for the inconvenience. Trivialist (talk) 20:55, 9 March 2025 (UTC)

@Trivialist: deactivated 506; this page is (probably) not monitored, if you add a ping to me I see the request and proceed quickly. Having a dedicated page for these requests would be useful (most are send directly in d:User talk:Magnus Manske presently), cf. my message at d:Wikidata talk:Mix'n'match#Procedure and criteria for requests of catalogues' deactivation if you want to comment. Epìdosis 11:15, 6 April 2025 (UTC)

APS Fellows instance of

https://mix-n-match.toolforge.org/?#/catalog/997 currently sets instance of (P31) physicist (Q169470) rather than instance of (P31) human (Q5) and occupation (P106) physicist (Q169470). AntiCompositeNumber (talk) 02:19, 31 March 2025 (UTC)

Maybe things like this could be requested to be changed via some integrated Catalog Update Requests set of functionalities (incl a table listing all update requests within the tool) as outlined below? Prototyperspective (talk) 18:06, 6 June 2025 (UTC)

Catalog Update Requests

Is there some page to request updates to existing catalogs? If not here's one request (and example):

• Subreddit Catalog for subreddit (P3984) – could somebody update it so that new subs are included and private subs and subs where posts are restricted to approved posters (generally few people and dead sub and not useful) are removed or moved to not applicable? The subs list seems to be from here. There probably is a straightforward way to get this list in some other format like via the reddit API or adding something to the URL and there could also be separate sites that have these. Alternatively one could maybe extract it from the HTML where the sub is at data-prefixed-name="…

I hope that if something goes wrong, one can easily revert the update to the prior state. I think there's probably some people pretty experienced in things like this and in this case I've already asked the catalog's creator. Prototyperspective (talk) 21:25, 4 June 2025 (UTC)

Is there a way to batch-add N/A tags?

Catalog #6816 (for GitLab.com topic ID (P13397)) has 263918 entries, but the vast majority of them are junk. Quite a lot of the junk follows predictable patterns, so I'm wondering if there's a way to do a batch job to mark things as N/A – it would make Match Mode a little bit more usable. Here are a few patterns I've identified so far that are always junk (there are surely many more):

• Prod_Live_Ver:.*
• Prod_Prev_Ver:.*
• appcon[a-z\d]{6,7}
• vo-product:.*
• stable-release:.*
• owner:.*
• db-owner=.*
• \d{5,}
• snow-ritm:ritm\d+

Just those first 3 regexes would eliminate many many thousands of junk items. Is there any way to do this? Thanks, IagoQnsi (talk) 19:09, 6 July 2025 (UTC)

I tried downloading the catalog, flagging all the junk rows in OpenRefine, and uploading them with the `q` column set to Q0, but that didn't work; MnM didn't actually mark them N/A. So instead I just wrote a quick Bash script to emulate clicking "N/A" on every single item one-by-one. Took a while to run (slept for 0.2 seconds between each request to avoid overwhelming the server), but it got the job done! –IagoQnsi (talk) 02:56, 15 July 2025 (UTC)

That sounds like something that could be very useful to others. If you made all this effort to try to solve this, identify what could be done, and writing the bash script, I think it would be great if you published it and documented in easy steps how to redo what you did so that other people with similar problems could also solve this. Maybe it could also be adjusted to for example remove all the preliminary matching to items that are scholarly papers or music albums etc in Subreddit Catalog. Prototyperspective (talk) 23:15, 25 July 2025 (UTC)

It wasn't anything complicated, just a Bash one-liner:

cat mnm-ids.txt | while read p; do curl 'https://mix-n-match.toolforge.org/api.php' --compressed -X POST -H 'Accept: */*' -H 'Content-Type: application/x-www-form-urlencoded; charset=UTF-8' -H 'Cookie: REDACTED' --data-raw "query=match_q&tusc_user=IagoQnsi&entry="$p"&q=0"; echo; echo $p; sleep 0.2; done

Replace "REDACTED" with the cookies from your browser, and "IagoQnsi" with your own Wikimedia username. Every line of mnm-ids.txt was the catalog entry ID (from the "#entry_id" column when you download an MnM catalog as a TSV) of an entry I wanted to mark N/A. I don't really want to make this any more polished or usable because it's really not ideal to hit MnM with thousands of requests – it would be better to update MnM to support Q0 in batch imports. IagoQnsi (talk) 20:38, 4 September 2025 (UTC)

Great, thanks! Would probably be best if this was added to some page about useful Mix-n-Match scripts or something like that (you could also set up such a page). If it's not long and complicated that's all the better! Do you think this could be adjusted to for example remove all the preliminary matching to items that are scholarly papers or music albums (a few other things too probably) in the preliminary matchings of the Subreddit Catalog? Because those are essentially all mismatched. I think this and also your script as is could be very useful for lots of MnM catalogs, thereby having potential to improve much WD data and freeing WD editor time who may do this more manually or who take longer for tasks such as matchings subreddits because there's so many items cluttering the page. Prototyperspective (talk) 19:50, 22 September 2025 (UTC)

Requesting catalog_editor access

Hi @Magnus Manske, I would like to become a catalog admin to update the Mix'n'match catalog I created along with future ones. Thanks! --Iketsi (talk) 18:22, 26 September 2025 (UTC)

Done. --Magnus Manske (talk) 11:10, 27 November 2025 (UTC)

How to add a newly created Wikidata item to an existing MnM entry?

I created catalog #7244 for the Nebraska Authors database a month or two ago, and the corresponding Wikidata property (P13862) was recently approved. How can I add this property to the existing catalog? I haven't had any luck using the "import or update catalog" page. Am I missing some permissions, or is there something else that I've missed?

Thanks!

DontCallMeLateForDinner (talk) 18:59, 20 October 2025 (UTC)

Done. You need indeed to have “catalog editor” permissions to do that − then it’s available at https://mix-n-match.toolforge.org/#/catalog_editor/7244 Jean-Fred (talk) 21:55, 20 October 2025 (UTC)

 This section is resolved and can be archived. If you disagree, replace this template with your comment. Prototyperspective (talk) 22:43, 26 November 2025 (UTC)

Is Mix'n'Match down ?

While trying to access https://mix-n-match.toolforge.org/ this morning, I just get a very blank screen (with only the top "Toolforge" icon)

So, is it M'n'M ? or Toolforge ?

Thanks for your help, Markus :) Hsarrazin (talk) 06:26, 28 November 2025 (UTC)

Date format in already imported set

I'was trying to import set of 1k polish novels – but I think I've made a mistake sending publication date in same format as in documentation for death and born fields. Now I cannot create new items from this set (clicking new item just sends request to api.php, gets answer with code 200, and message Bad value type quantity, expected time (that's why I'm thinking the cause is format of publication date, but I'm not sure), and keeps trying this all over again sending same request).

Is there possibility to upload data with changed format? Because when I've tried it Mix'n'matched just add another column for publication date (which make sense in some cases, but is not what I need).

Or maybe I should remove set and reupload it? If yes – how (cannot find this option)?

thanks for help in advance, regards, Piastu βy język giętki… 10:50, 6 January 2026 (UTC)