Wikidata:Property proposal/South African Company Registration Number

South African Company Registration Number

Originally proposed at Wikidata:Property proposal/Authority control

Done: South African Company Registration Number (P8182) (Talk and documentation)

Description	Registration number of South African companies issued by the Companies and Intellectual Property Commission of South Africa
Represents	South African Company Registration Number (Q91317879)
Data type	External identifier
Domain	instance of (P31)company (Q783794) and country (P17)South Africa (Q258)
Allowed values	`\d{4}/\d{6}/\d{2}`
Example 1	Shoprite (Q1857639) → 1936/007721/06
Example 2	ABSA Group Limited (Q58641733) → 1986/003934/06
Example 3	Old Mutual (Q289704) → 2017/235138/06
Example 4	Telkom (Q1818970) → 1991/005476/30
Source	http://eservices.cipc.co.za/Search.aspx
Planned use	I want to be able to set this for South African companies when I update their details. At the moment I only wanted it in cases where I was updating South African companies by hand but it may be useful for a crawler of Legal Entity Identifier (Q6517388) which should set this for any entity that is registered at South Africa's Companies and Intellectual Property Commission (Q39086117) (RA000531) and these identifiers are also used by OpenCorporates (Q7095760)(more info). This identifier should ideally be set for all items corresponding to South African countries but it should also not be considered a property that indicates notoriety.
Number of IDs in source	More than 1 million, should not be more than 25 million.
Expected completeness	always incomplete (Q21873886)
See also	Norwegian organisation number (P2333) Swedish Organization Number (P6460) KvK company ID (P3220) Swiss Enterprise Identification Number (P4829) OpenCorporates ID (P1320) Legal Entity Identifier (P1278)

Motivation

Add registration numbers for south African Companies. Sadly I am not aware of any site that can be used as formatter url. Iwan.Aucamp (talk) 13:02, 9 April 2020 (UTC)[reply]

More info:

Issued by: South Africa's Companies and Intellectual Property Commission (Q39086117)
details on format
Search sites:
- Companies and Intellectual Property Commission - eService / Search
- Biz Portal/Biz Profile
Other databases referencing this:
- https://opencorporates.com/
- https://www.gleif.org/
LEI Registration Authority Code RA000531

Iwan.Aucamp (talk) 08:58, 21 April 2020 (UTC)[reply]

Discussion

Notified participants of WikiProject Companies Jklamo (talk) 21:09, 19 April 2020 (UTC)[reply]

Notified participants of WikiProject Authority control Iwan.Aucamp (talk) 01:16, 21 April 2020 (UTC)[reply]

Support ArthurPSmith (talk) 18:34, 9 April 2020 (UTC)[reply]
Support --Jklamo (talk) 21:11, 19 April 2020 (UTC)[reply]
Support https://opencorporates.com/registers/237 says that register is closed so opencorporates cannot crawl them. Then I think we don't need them either --Vladimir Alexiev (talk) 13:58, 20 April 2020 (UTC)[reply]
- @Vladimir Alexiev: South African Company Registration Number is still in use and relevant. Why does it matters if the database can be crawled or not? Opencorporates can be crawled and the information for other notable items could be added manually. This would not be a property that signifies notoriety anyway. Iwan.Aucamp (talk) 22:16, 20 April 2020 (UTC)[reply]
- @Iwan.Aucamp: The OpenCorporates data is 5 years old, so I still don't care much about this data if there's no up to date site where it can be checked. Here is some info for the proposal:
  - Formatter URL: https://opencorporates.com/companies/za/$1
  - Regex: \d{4}-\d{6}-\d{2}, eg https://opencorporates.com/companies/za/6752-752049-07. Maybe the new delimiter is slash, but the first number can start with any digit not just "1|2"
  - Number of entries: 1,559,848; point in time: Sep 2014; reference https://opencorporates.com/companies/za, https://opencorporates.com/registers/237.
    - 783,614: applies to part: In Business (active); same qualifiers --Vladimir Alexiev (talk) 06:59, 21 April 2020 (UTC)[reply]
  - @Vladimir Alexiev: There are sites where it can be checked: Companies and Intellectual Property Commission - eServices and Biz Portal/Biz Profile. The limitation is just that the terms and conditions of these sites prohibit "crawling" without prior permission. It very well may be allowed on request. If there is a restriction on the creation of properties for authority control that cannot be crawled or which have no formatter URL then I can understand objections, but that is not the case as far as I know. I changed the regex back to what it when I created the proposal, the format was never with dashes, it was always with slashes: https://eservices1.capetown.gov.za/sap/bc/webdynpro/sap/zcur_company_reg_hint# (this list may not be exhaustive) Iwan.Aucamp (talk) 07:46, 21 April 2020 (UTC)[reply]
  - @Iwan.Aucamp: Ok then, changed to "Support". Change the URL to \d{4}/\d{6}/\d{2}: this shows exactly where the dashes should be. @Pigsonthewing: there is a way to translate slashes to dashes for the OCorp "third party formatter", right? If OCorp examples are correct, the first 4 digits are not YYYY. Iwan, please state your plans what you'd do with the prop once it's created. Cheers! --Vladimir Alexiev (talk) 15:51, 21 April 2020 (UTC)[reply]
    - @Vladimir Alexiev: Thanks, I have updated the allowed values to \d{4}/\d{6}/\d{2} and added some details to planned use. Iwan.Aucamp (talk) 18:37, 21 April 2020 (UTC)[reply]
    - "Formatter URL: https://opencorporates.com/companies/za/$" that would be "Third-party formatter URL". Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 08:57, 21 April 2020 (UTC)[reply]
- Openness (or lack of it) is not a good reason. Apparently number is still in use, just cannot be mass crawled. Linking individual entries is OK, we can use GLEIF dataset as well.--Jklamo (talk) 08:01, 21 April 2020 (UTC)[reply]
Support. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 08:56, 21 April 2020 (UTC)[reply]
I think information about the issuing organization should be in the property description. ChristianKl ❪✉❫ 15:34, 21 April 2020 (UTC)[reply]
- @ChristianKl: I added it. It is also on Wikidata item of this property (P1629)South African Company Registration Number (Q91317879) and should be added to the property when created as usual Iwan.Aucamp (talk) 18:04, 21 April 2020 (UTC)[reply]
An obvious constraint for the Domain line: country (P17) South Africa (Q258). P17 is required by OpenCorporates ID (P1320) and I have added it to Shoprite (Q1857639) as an example --Vladimir Alexiev (talk) 16:03, 21 April 2020 (UTC)[reply]
- @Vladimir Alexiev: Added Iwan.Aucamp (talk) 18:42, 21 April 2020 (UTC)[reply]
@Iwan.Aucamp, ArthurPSmith, Jklamo, Vladimir Alexiev, Pigsonthewing, ChristianKl: Done --Tinker Bell ★ ♥ 05:38, 4 May 2020 (UTC)[reply]

I fixed teh regex as described above (to start with any digit) and added Number of Records. Didn't add "third party formatter" because there's a problem with slash vs dash. --Vladimir Alexiev (talk) 07:58, 29 May 2020 (UTC)[reply]