Closed ThanePi closed 1 year ago
Hi @ThanePi ,
Why do we need to separate vocab
from cdm
?
Hi @ThanePi ,
Why do we need to separate
vocab
fromcdm
?
Hey @na399,
It's for scalability purposes, I found that most repo in OHDSI (such as ACHILLES used in ATLAS) have implemented parameterized schemas for "vocab", "cdm" and "result". Utilizing this should allows us for greater flexibility in further scaling (example: Having just single vocab server shared for all org, instated of each org ran their own vocab). Additionally, I believed that OHDSI team has divided their work into separate modules/teams to contribute to the overall OMOP project independently, this also for scaling purpose.
Another reason is that on early development of hosting ATLAS on SiData Warehouse, I also split OMOP DB into 3+ schema follows their permission by use case ("cdm" is read only, "vocab" is read only and write only for R server, "results" and "temp" are both read and write)
btw, We can discuss if this approach is match for our use case or to discuss further with our consultants.
Add and rework on some configuration to support Vocabulary in terms of
omop
also add somevocab
constraint test case.omop-pg
now includevocab
schema and its DDLcdm
andvocab
instated of singleomop