Missing support for GBIF triplet style ID

[thomasstjerne]

Transferred from gbif/portal-feedback#588

Missing support for GBIF triplet style ID

This dataset from Togo uses the GBIF triplet style identifier, which is a combination of three fields: catalogNumber, institutionCode and collectionCode.

Each of its records is therefore uniquely identified.

Contrary to the validator report, the file should be indexible by GBIF and each record should be uniquely identified.

Of course we want to promote the use of occurrenceIDs instead of the GBIF triplet, however, the validator is supposed to reflect GBIF indexing operating procedures, which supports the GBIF triplet identifier afaik.

fbitem-9ceba7755c086fa5a18ef4c6bd2102ecccab59e7
User provided contact info: kbraak
System: Chrome 61.0.3163 / Mac OS X 10.10.5
Referer: https://www.gbif.org/tools/data-validator/1508848348041
Window size: width 1338 - height 843
API log
Site log

[kbraak]

Supposedly the above dataset cannot be indexed at GBIF because 29 out of 8015 records have duplicate triplets (see below - thanks @jlegind for doing the analysis). If this is true, the validator should explain this precisely and clearly - currently the user isn't warned of these duplicates.

institutionCode collectionCode catalogNumber Count
Herbarium togoense TOGO 9809 2
Herbarium togoense TOGO 1458 2
Herbarium togoense TOGO 910 2
Herbarium togoense TOGO 11359 2
Herbarium togoense TOGO 28 2
Herbarium togoense TOGO 2131 2
Herbarium togoense TOGO 1053 2
Herbarium togoense TOGO 2744 2
Herbarium togoense TOGO 11874 2
Herbarium togoense TOGO 7323 2
Herbarium togoense TOGO 8376 2
Herbarium togoense TOGO 1387 2
Herbarium togoense TOGO 11217 2
Herbarium togoense TOGO 3751 2
Herbarium togoense TOGO 11587 2
Herbarium togoense TOGO 10972 2
Herbarium togoense TOGO 2567 2
Herbarium togoense TOGO 2566 2
Herbarium togoense TOGO 9384 2
Herbarium togoense TOGO 1345 2
Herbarium togoense TOGO 2568 2
Herbarium togoense TOGO 4928 2
Herbarium togoense TOGO 7131 2
Herbarium togoense TOGO 1675 2
Herbarium togoense TOGO 11004 2
Herbarium togoense TOGO 12113 2
Herbarium togoense TOGO 7926 2
Herbarium togoense TOGO 10777 2
Herbarium togoense TOGO 11226 2

[cgendreau]

Triplet are not supported by purpose and all new datasets shall use a single identifier field (e.g. occurrenceId).

[cgendreau]

ah ok I see, the validation result includes nothing except empty recordId.
Yes, that's indeed a problem.