OHDSI Home | Forums | Wiki | Github

ETL-Synthea: Instructions to create CDM Postgres Database missing?

(Sam Habiel) #1


As I mentioned in the introduction in another thread, I am interested in putting together the OHDSI software on my computer on my own, without the use of docker or AWS images.

I am reading this in the repo.

 # We are loading into a local PostgreSQL database called "synthea10".  
 # The schema to load the Synthea tables is "native".
 # The schema to load the Vocabulary and CDM tables is "cdm_synthea10".  
 # The username and pw are "postgres" and "lollipop".
 # The Synthea and Vocabulary CSV files are located in /tmp/synthea/output/csv and /tmp/Vocabulary_20181119, respectively.

But I don’t see any instructions on creating the CDM database on postgres, or whether it is even necessary. I hope somebody can enlighten me.

I also read the first few chapters of the OHDSI book, and I couldn’t find any links to instructions on setting up an empty CDM database.



Hi Sam,

The CDM Postgres scripts are here:

Step 3 in the README includes loading the vocabulary as well as your data. You can choose and download the vocabulary from https://athena.ohdsi.org. You would need a UMLS account to put the CPT4 codes together before importing (instructions included in download.)

Hope this helps!

(Sam Habiel) #3


I am writing back to update everybody on what I found out.

I had difficulty finding a branch that works, and I finally found one here (https://github.com/Roeder-OHDSI-Forks/ETL-Synthea-1/tree/v5.3.1-updates-maintenance).

There are multiple issues right now, but these are for another thread.

Short answer is that the CDM database gets populated using scripts in this repository; however, you need to create the original database and the two required schemas in Postgres before you start.

You can do that by:

CREATE SCHEMA cdm_synthea10;

At long last, I finally have a v5.3.1 CDM database.