[06:52:56] hi everyone [06:53:16] about bug #42659, a SMW dev could help. [06:54:29] The query for getting all indizes filters out all indizes created with "UNIQUE INDEX", but for smw_prop_stats such an index is added explicitly [06:55:04] resulting in SMW_setup trying to recreate that index since it doesn't show up [06:55:33] ... in getIndexInfo [06:56:24] any suggestions on how to work around this? [06:59:39] one possibility would be to add another loop before "// Add new indexes" (SMW_SQLHelpers.php) which goes through $indexes and looks if one is a UNIQUE INDEX and in that case do another lookup and based on that unset $indexes[$id] [06:59:43] would that be a solution? [09:16:10] well, after some more digging in the code: the answer is to skip the creation of a UNIQUE INDEX since p_id is defined as a PK which is equal to UNIQUE NOT NULL and columns with the UNIQUE constraint get an implicit UNIQUE INDEX on PostgreSQL [10:32:22] anybody have any idea what the dynatree part of Semantic Forms is there for? [13:16:52] <_ken_> I have a question: [13:17:08] <_ken_> So I understand a bit about inline queries via #ask and #show [13:17:38] <_ken_> My understanding is that these return the page and/or the value of the property depending on the query [13:17:57] <_ken_> I understand how to impliment that [13:18:20] <_ken_> My issue is, what if a page has two property values for the same property? [13:18:43] <_ken_> Like say my page has [[Whatevervalue::x]] and [[Whatevervalue::y]] [13:20:08] <_ken_> If my query is {{#ask: [Category:MyCat] | ?Whatevervalue }} I end up getting a column of Whatevervalue but it's listed as x,y [13:21:00] <_ken_> specifying limit= (such as {{#ask: [Category:MyCat] | ?Whatevervalue | limit=1}} doesn't help as it appears that parameter is used to specify the number of pages returned, not the number of properties [13:21:15] <_ken_> Is there a mechanism to seperate my x and y into seperate columns? [13:22:40] <_ken_> I know when I browse properties by name, it also shows my property as "x,y" (minus quotes), but I expect I can make a query to seperate these? [13:26:46] (PS1) Dev-zero: Fix some incompatibilities with PostgreSQL [extensions/SemanticMediaWiki] - https://gerrit.wikimedia.org/r/88459 [13:33:07] (PS2) Dev-zero: Fix some incompatibilities with PostgreSQL [extensions/SemanticMediaWiki] - https://gerrit.wikimedia.org/r/88459 [13:36:41] (CR) Jeroen De Dauw: "(3 comments)" [extensions/SemanticMediaWiki] - https://gerrit.wikimedia.org/r/88459 (owner: Dev-zero) [13:36:46] (CR) Jeroen De Dauw: [C: -1] Fix some incompatibilities with PostgreSQL [extensions/SemanticMediaWiki] - https://gerrit.wikimedia.org/r/88459 (owner: Dev-zero) [13:48:03] <_ken_> http://semantic-mediawiki.org/wiki/Help:Selecting_pages says: If a property that is used for sorting has more than one value for some page, then this page will still appear only once in the result list. The position that the page takes in this case is not defined by SMW and may correspond to either of the property values. In the above examples, this would occur if one city would have multiple population numbers specified, or if o [13:48:03] <_ken_> countries each of which has a population. It is suggested to avoid such situations. [13:48:18] <_ken_> That last sentence is disturbing. [13:49:00] <_ken_> Does it mean we're not meant to have pages with properties that have multiple values? [13:50:27] (CR) Mwjames: "Even though, this isn't part of this change, it would be great if instead pulling dependencies ($wgDBtype == 'postgres') into the Store an" [extensions/SemanticMediaWiki] - https://gerrit.wikimedia.org/r/88459 (owner: Dev-zero) [14:16:38] _ken_: the handling of mutliple values for a single page is not ideal, but it's certainly not discouraged to have that. [14:18:17] That text you quoted, as far as I can tell, only refers to properties that are used for sorting. [14:38:26] (CR) Dev-zero: "(didn't figure out how to display and respond per-commit changes in gerrit)" [extensions/SemanticMediaWiki] - https://gerrit.wikimedia.org/r/88459 (owner: Dev-zero) [14:40:44] JeroenDeDauw: I'm here, btw, if you want to discuss [14:40:44] <_ken_> Yes you are correct it handles the sorting. The data always comes out in the format: Page then properties (which would be something like MyPageName then x,y) regardless of whether I strip the title field, or leading field, etc. [14:41:12] <_ken_> It seems I'm just going to use the #explode function to seperate the data into it's subparts (to remove the comma) [14:42:00] <_ken_> I expected I could output the individual property values via a query and it seems I can't when they are on the same page [14:43:07] <_ken_> I mean, what if we had [[Eye color::blue]] [[Eye color::black]] to describe a single page (say of type Person) [14:43:27] <_ken_> Somone can have different color eyes and we should be able to have a query that seperates the two [14:43:32] <_ken_> Maybe that's a poor example [14:43:41] <_ken_> #explode it is, I guess [14:47:56] (CR) Jeroen De Dauw: "James, agree the store is a mess, though this is clearly not in scope of this commit as you mentioned. We could clean up the setup and sch" [extensions/SemanticMediaWiki] - https://gerrit.wikimedia.org/r/88459 (owner: Dev-zero) [14:50:12] dev-zero: it is rather hard for me to verify this change... [14:52:57] (CR) Jeroen De Dauw: "(2 comments)" [extensions/SemanticMediaWiki] - https://gerrit.wikimedia.org/r/88459 (owner: Dev-zero) [14:56:40] dev-zero: https://gerrit.wikimedia.org/r/#/c/88459/1/includes/storage/SMW_SQLHelpers.php [14:56:53] dev-zero: also notice the link to the SetupHandlers file at the top right [14:57:03] dev-zero: I added your comments there already now [14:57:11] Will reply in a moment [15:01:25] _ken_: you should use #arraymap, if I understand the issue correctly. [15:01:49] ...but I thought that you already had the multiple values being stored, and you were just wondering how best to display them in a query. [15:08:09] (CR) Jeroen De Dauw: "(2 comments)" [extensions/SemanticMediaWiki] - https://gerrit.wikimedia.org/r/88459 (owner: Dev-zero) [15:08:13] dev-zero: there you go [15:37:15] <_ken_> yaron: Yeah I already have multiple values stored. #explode is working fine I hadn't known about it until today. Thanks again [15:37:55] Alright then. [15:38:16] #arraymap is better. [15:43:58] JeroenDeDauw: thanks, get it now [15:48:19] (PS1) Yaron Koren: New version: 2.6 [extensions/SemanticForms] - https://gerrit.wikimedia.org/r/88487 [15:51:41] (CR) Yaron Koren: [C: 2 V: 1] New version: 2.6 [extensions/SemanticForms] - https://gerrit.wikimedia.org/r/88487 (owner: Yaron Koren) [15:51:43] (Merged) jenkins-bot: New version: 2.6 [extensions/SemanticForms] - https://gerrit.wikimedia.org/r/88487 (owner: Yaron Koren) [15:53:09] JeroenDeDauw: wrt the changes which change the name of the indices: those should work for all databases, the only problem I see with them is if a DBMS has restrictions on the length of the name [15:54:19] JeroenDeDauw: HUNK 3 for SMW_SQLHelpers.php is really just a bugfix [15:55:39] JeroenDeDauw: wrt the quotation problem of the tablename: I can check the other solution mentioned in the comment [15:56:06] JeroenDeDauw: ... would be nicer since there is postgres-specific code in getIndexInfo already [16:23:40] <_ken_> yaron, I will look into using #arraymap then. Just glancing at the manual for it right now it seems that will work for me. I had previously thought it was a property type (thus I thought I would have to drastically change my layout). [18:38:30] (PS1) Mwjames: Make sure that only SQLStore tests run on a SQLStore [extensions/SemanticMediaWiki] - https://gerrit.wikimedia.org/r/88530 [18:39:58] (CR) Jeroen De Dauw: [C: 2] Make sure that only SQLStore tests run on a SQLStore [extensions/SemanticMediaWiki] - https://gerrit.wikimedia.org/r/88530 (owner: Mwjames) [18:40:03] (PS2) Jeroen De Dauw: Make sure that only SQLStore tests run on a SQLStore [extensions/SemanticMediaWiki] - https://gerrit.wikimedia.org/r/88530 (owner: Mwjames) [18:40:08] (CR) Jeroen De Dauw: [C: 2] Make sure that only SQLStore tests run on a SQLStore [extensions/SemanticMediaWiki] - https://gerrit.wikimedia.org/r/88530 (owner: Mwjames) [18:41:52] (CR) jenkins-bot: [V: -1] Make sure that only SQLStore tests run on a SQLStore [extensions/SemanticMediaWiki] - https://gerrit.wikimedia.org/r/88530 (owner: Mwjames) [18:42:51] dev-zero: can you please reply inline? Just click one of my comments and then hit the reply button (which is only there if you are logged in) [18:42:59] Very hard to keep track of this now [18:43:26] (CR) Mwjames: "It wasn't me ..." [extensions/SemanticMediaWiki] - https://gerrit.wikimedia.org/r/88530 (owner: Mwjames) [18:43:27] (Merged) jenkins-bot: Make sure that only SQLStore tests run on a SQLStore [extensions/SemanticMediaWiki] - https://gerrit.wikimedia.org/r/88530 (owner: Mwjames) [18:46:33] (PS1) Mwjames: [WIP] [DNM] SMW\MongoStore, using MongoDB as Sementic Mediawiki storage backend [extensions/SemanticMediaWiki] - https://gerrit.wikimedia.org/r/88534 [18:54:49] (CR) Mwjames: "$smwgDefaultStore = "\SMW\MongoStore\DocumentStore";" [extensions/SemanticMediaWiki] - https://gerrit.wikimedia.org/r/88534 (owner: Mwjames) [19:39:41] (PS1) Yaron Koren: Updated for version 2.6 [extensions/SemanticForms] - https://gerrit.wikimedia.org/r/88617 [19:44:38] (CR) Yaron Koren: [C: 2 V: 1] Updated for version 2.6 [extensions/SemanticForms] - https://gerrit.wikimedia.org/r/88617 (owner: Yaron Koren) [19:45:48] (CR) Yaron Koren: [V: 2] Updated for version 2.6 [extensions/SemanticForms] - https://gerrit.wikimedia.org/r/88617 (owner: Yaron Koren)