| OHDSI Home | Forums | Wiki | Github |
I wouldn’t think the constraints should matter, but agreed on the indexes. Depending on the hardware behind the database, Achilles can take many hours to finish.
I have obtained an Oracle instance to test against, working with our IT group to get a CDM landed on it. Once that is available, I’ll be able to test the oraclefixes branch more thoroughly.
Constraints would have no affect on performance of the query. I’d check to see the load on the Oracle server with a DBA to see if it could be tuned further. I am unaware if there are best practices in further performance tuning of the CDM in Oracle however. An Oracle DBA should have suggestions if OHDSI community does not have them though.
Hi @Ajit_Londhe, @Dave.Barman, @Chris_Knoll,
I was able to successfully complete the Achilles run with my test data and I am also able to see the results in Atlas.
It did take around 6 hrs to complete, but I am happy that the data is visible now in Atlas.
I will be reviewing the results in Atlas now and will get back to you for any issues.
Thanks a lot for your help and support and being patient with me. I was about to give up, but the support I got from you folks did not let me do that. Thanks again.
@Ajit_Londhe - FYI: I was able to complete the process by running the achillesHeel R command directly while passing oracleTempSchema to the call. So, I maybe the main achilles R call just needs to be modified to pass the oracleTempSchema variable if the DBMS is ‘oracle’ …
Hi @rkboyce,
Hmm, it does send over the oracleTempSchema. My last commit to the branch was 26 days ago, do you know if your installation has that commit?
An update on our Oracle test instance: we have the instance, but working with our IT group on permissions. I expect to have that resolved by Wednesday at the latest, and then we have a DBA that can land a CDM there pretty quickly thereafter. Hoping to get this validated against Oracle by end of week.
Thanks,
Ajit
No problem @smanikonda! Sorry it’s taken a while to get this resolved, but I think soon we’ll have a good way to validate Achilles against Oracle, which has been a blind spot.
Just to say, we are testing our new psql database. We are also stuck on test 117 for no apparent reason.
It doesn’t throw out a error as it is stuck there. So there is no error log to rely on too.
I am curious what is so special about 117 that most of the test runs fast before it and then it is stuck in 117.