Stream: implementers
Topic: Any servers implemented the $stats operation?
Eric Haas (Mar 10 2020 at 04:55):
I've been reviewing the $stats operation for a simple use case: calculate patient reported activity data from FHIR Observations: e.g. average, min, max, count for step count. However, it seems that this operation has not been implemented on the reference servers. Has anybody implemented it besides GG?
Is there any inherent advantage to an operation like this vs flattening the data and using dataframes for these simple statistics?
Grahame Grieve (Mar 10 2020 at 06:15):
flattening the data and using dataframes
what does that mean?
Abel Stolz (Mar 10 2020 at 06:21):
"flattening the data" means writing your data into a csv file. Dataframes are used in R and are essentially tables, as are data in csv format, so I suppose the intent is to do statistics on the data.
Grahame Grieve (Mar 10 2020 at 06:51):
how is that meaningfully different to the $stats operation?
Eric Haas (Mar 10 2020 at 18:24):
i agree, behind the scenes it may be the same, but the question I have is whether the $stats operation is getting any traction? if not - why not?
Grahame Grieve (Mar 10 2020 at 19:16):
no sure why not. I think it matters
Grahame Grieve (Mar 10 2020 at 19:16):
but it is hard
nicola (RIO/SS) (Mar 18 2020 at 17:15):
We decided that SQL on FHIR is a more generic solution for these problems
nicola (RIO/SS) (Mar 18 2020 at 17:16):
And just provide sql endpoint on our server - https://docs.aidbox.app/advanced/usdpsql and https://docs.aidbox.app/basic-concepts/search-1/custom-search
nicola (RIO/SS) (Mar 18 2020 at 17:18):
Here is how you can turn SQL into csv - https://docs.aidbox.app/basic-concepts/bulk-api-1/usddump-sql-tutorial
Last updated: Apr 12 2022 at 19:14 UTC
