Skip to content

Comments

Payer mapping dups 5263#78

Open
schandaka wants to merge 588 commits intoARCH-commons:masterfrom
kumc-bmi:payer_mapping_dups_5263
Open

Payer mapping dups 5263#78
schandaka wants to merge 588 commits intoARCH-commons:masterfrom
kumc-bmi:payer_mapping_dups_5263

Conversation

@schandaka
Copy link

No description provided.

dckc and others added 30 commits April 14, 2017 14:57
medication_id_to_best_rxcui: include curated mappings
and un-hard-code ETL schema
  - compute children_loinc_codes directly without
    first computing parent_loinc_codes
  - use c_hlevel to cut down parent / child search space
  - factor out
    - lab_terms: where c_fullname like '\i2b2\Laboratory Tests\%'
    - loinc_terms: where c_basecode like 'LOINC:%'
  - use just 3 columns for the tricky parent/child part
    - Oracle seems to build a temp table for this

Fetched 50 rows in 122.732 seconds
jgaupp and others added 30 commits July 20, 2018 13:56
The encounter query was running for more than 6 hours in B2, versus
just under 2 hours in A1.  The enctype subquery seemed to be the
contributing factor.
The real performance hit came on the observation_fact join.  Run time
on B2 jumped from 15 minutes to 15 hours.  Isolating ob_fact reduced
runtime to 1 hour.
This could be moved back to cdm_prep in Heron code.  Not sure what's
the best fit.
For a proper fix, the pcornet_vital.concept_cd provided by
scilhs-ontology will need to be updated.
Level of effort proved to be minimal despite novel work in H2P and
SCHILS.  Tested okay in STAGEDEV.
The dependency is needed for a join on encounter to eliminate med_admin
facts outside the timeline required for CDM.
Fixed incorrect join syntax
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

5 participants