Missing primary key definition for Cohort table in Postgres

[sameertj]

The OMOPCDM_postgresql_5.4_primary_keys.sql script for CDM 5.4 does not define a primary key for the Cohort table. This breaks the OMOPCDM_postgresql_5.4_constraints.sql script.

The fix is to add the following to the OMOPCDM_postgresql_5.4_constraints.sql script:

ALTER TABLE @cdmDatabaseSchema.COHORT ADD CONSTRAINT xpk_COHORT PRIMARY KEY (cohort_definition_id);

[motey]

Regarding the docs this table has no primary key https://ohdsi.github.io/CommonDataModel/cdm54.html#cohort

There are multiple tables with no PK. cohort cdm_source cohort_definition concept_ancestor concept_synonym drug_strength fact_relationship source_to_concept_map concept_relationship death episode_event

It is not prohibited to create tables without PKs, but it is at least awkward. from a database-operator and developer perspective, it makes some common tasks unnecessary complicated. I can not see the benefit, we get from the non-extisting PKs. Does anyone can provide the rational behind this design decision? I could not find anything. If there is none, i would highly be in favor of having a primary key for every table in future CDM versions. :heart:

I can create a PR if favored.