Wikidata:Property proposal/Riksdagen person guid
Riksdagen person guid
editOriginally proposed at Wikidata:Property proposal/Person
Description | identifier for a Riksdagen member or other person on riksdagen.se with format UUID |
---|---|
Represents | Swedish Parliament database (Q21592569) |
Data type | External identifier |
Domain | human (Q5), primarily members of Riksdagen, Swedish MEP |
Allowed values | [\da-f]{8}-[\da-f]{4}-[\da-f]{4}-[\da-f]{4}-[\da-f]{12} (Universally Unique Identifier (Q195284)) |
Example 1 | Stefan Löfven (Q2740012) → ac737989-5fa0-44bc-ad69-c1a0ddba71bb |
Example 2 | Carl Bildt (Q52922) → d7c317f1-83e4-11d4-ae60-0050040c9b55 |
Example 3 | Fredrik Reinfeldt (Q52920) → d7c31d61-83e4-11d4-ae60-0050040c9b55 |
Source | http://data.riksdagen.se/personlista |
Planned use | its used today in Riksdagen person-ID (P1214) and will be used in the future in the same way |
Number of IDs in source | its all Swedish Parlament members since 1971, ca. 5000 people |
Expected completeness | 100% |
Formatter URL | http://data.riksdagen.se/personlista/?iid=$1&utformat=html |
See also | Riksdagen person-ID (P1214) with format \d{10,13} |
Motivation
editthe Swedish Government are changing identifiers to a guid and told us to update Wikidata see task T235521. My intention is to convert the old ids to the new using the old property Riksdagen person-ID (P1214) see alsoT235521. If that is not ok we need a new property thats why this request...
I Oppose creating this property and suggest that we should just change values on Riksdagen person-ID (P1214) (see T235521 and discussion Property_talk:P1214) but user @Jura1: insits so lets take the discussion here. Let me know what you decide.
@Larske, Yger: do you have a second opinion - Salgo60 (talk) 00:22, 24 November 2019 (UTC)
- Support I changed my mind because we found a CSV file with the old IDs so its better have two separate properties in WIkidata one for the old one for the new... see video in swedish - Salgo60 (talk) 14:34, 22 June 2020 (UTC)
Discussion
edit- Comment Lavallen set up Riksdagen person-ID (P1214) five years ago and this has since been made available here. As most of his contributions these have proven to be sound. The data isn't "wrong" and shouldn't be deleted, merely because the original website is no longer available. --- Jura 00:27, 24 November 2019 (UTC)
- The original website is available and will be so I guess as long as Sweden (Q34) exists (see xml with oldid in <intressent_id> ? 0622185335615 and new id in <sourceid> d7c317f1-83e4-11d4-ae60-0050040c9b55 ), but the original website had a technical debt using short ids and is now moving into guids... they will support both ids for the next 2-3 years. For me it doesnt matter more than I have to write this proposal and wait some weeks for decision. Maybe we need to change some templates using WD properties...- Salgo60 (talk) 00:39, 24 November 2019 (UTC)
- Notified participants of WikiProject Data Quality do you have any philosophy/opinions if we should always create new identifiers when a system is changing identifiers - Salgo60 (talk) 21:29, 29 November 2019 (UTC)
- Seems to me a relatively pragmatic issue: Is there a stakeholder that benefits from retaining the old IDs? If yes they should obviously stay, if no, I would see it as sensible to drop/overwrite them, to avoid future confusions, and not clutter Wikidata with irrelevant content. Ls1g (talk) 10:28, 30 November 2019 (UTC)
- I agree. Adding a value with the new GUID makes sense. And we should set the other value as deprecated. Ainali (talk) 13:41, 1 December 2019 (UTC)
- Unless the identifier was irrelevant in the first place, I think it should be preserved. Users who did use it can continue to use it to find equivalents or lookup pages in the internet archive. We don't delete or overwrite identifiers because Google or some government takes content offline. --- Jura 17:08, 1 December 2019 (UTC)
- I totally agree with the comments mentioned above, and using the rank to signal the validity of the values is very useful. Cristina Sarasua (talk) 10:11, 10 January 2020 (UTC)
- Totally agree. We just need to bear in mind that Help:Ranking#Deprecated_rank doesn't support deprecation in the case here. --- Jura 10:11, 20 January 2020 (UTC)
- I totally agree with the comments mentioned above, and using the rank to signal the validity of the values is very useful. Cristina Sarasua (talk) 10:11, 10 January 2020 (UTC)
- Unless the identifier was irrelevant in the first place, I think it should be preserved. Users who did use it can continue to use it to find equivalents or lookup pages in the internet archive. We don't delete or overwrite identifiers because Google or some government takes content offline. --- Jura 17:08, 1 December 2019 (UTC)
- I agree. Adding a value with the new GUID makes sense. And we should set the other value as deprecated. Ainali (talk) 13:41, 1 December 2019 (UTC)
- Support --- Jura 16:12, 19 April 2020 (UTC)
- Support --- Popperipopp (talk) 15:17, 12 June 2020 (UTC)
- Comment --- What is the status of this discussion? I tried to merge a dataset from Riksdagen with Wikidata but ended up with lots of non-matches. Currently the identifier situation seem stuck between the "old" and "new" identifiers, which is probably the worst case scenario. Popperipopp (talk) 14:58, 12 June 2020 (UTC)
- PS. Note that the "new" identifier proposed on this page is not even present in the data dumps of members of parliament from data.riksdagen.se. Popperipopp (talk) 15:01, 12 June 2020 (UTC)
- @Popperipopp: the proposal needs additional support. Not sure why @Criscod, Ls1g, Ainali: want us to drop the identifier present in the data dumps. --- Jura 15:04, 12 June 2020 (UTC)
- I was thinking of adding the annotation that indicates whether it should still be used or not (e.g., with a rank), not really dropping the identifier. Cristina Sarasua (talk) 08:00, 15 June 2020 (UTC)
- Even though I personally would prefer having a distinct identifier, I think your proposal makes sense as well. What does not make sense is simply replacing current identifiers with new ones, so I'm adding the deleted identifiers back in, in the meantime. Popperipopp (talk) 12:34, 15 June 2020 (UTC)
- I was thinking of adding the annotation that indicates whether it should still be used or not (e.g., with a rank), not really dropping the identifier. Cristina Sarasua (talk) 08:00, 15 June 2020 (UTC)
Support Querying and maintaining distinctions based on rank gets messy (e.g. still cannot be set in QS. Belteshassar (talk) 14:47, 22 June 2020 (UTC)
- @Salgo60, Jura1, Ls1g, Ainali, Criscod, Popperipopp: @Belteshassar: Swedish Parliament person GUID (P8388) has been created. Pamputt (talk) 18:46, 27 June 2020 (UTC)
- @Belteshassar: did you mean filtering based on rank in a SPARQL query? in case it helps: https://en.wikibooks.org/wiki/SPARQL/WIKIDATA_Qualifiers,_References_and_Ranks#Ranks Cristina Sarasua (talk) 17:28, 11 July 2020 (UTC)
- Thanks @Criscod:. I was a bit imprecise above. What I meant is that at some point, if the old style ids are discontinued, we may have a situation where we have items with only new style, only old style or both ids. We can’t really expect anyone who creates an item for a new person to be aware that the id they add should be of preferred rank, so probably the only safe way to query for only new style or only old style ids would be to use a regular expression filter. Anyway, the new property seems to have reached consensus, so we are all good. Belteshassar (talk) 07:32, 12 July 2020 (UTC)
- @Belteshassar: did you mean filtering based on rank in a SPARQL query? in case it helps: https://en.wikibooks.org/wiki/SPARQL/WIKIDATA_Qualifiers,_References_and_Ranks#Ranks Cristina Sarasua (talk) 17:28, 11 July 2020 (UTC)