Page History
...
Info |
---|
i2b2 users must have the DATA_AGG user permission to view the counts through the web client. The stored procedures loaded in the Metadata schema must have read access to the CRC schema (more information in Installation below). Mapping codes through the concept_dimension or Adapter Mappings files are not supported. i.e. the c_basecodes in your ontology tables must be the same codes used in your fact tables. |
Installation
This should have already occurred in previous sections of this guide, but verify you have run these steps:
...
i2b2 includes the totalnum stored procedures and they are installed as part of the data installation. To manually install/upgrade, follow these steps:
- Download the i2b2 upgrade package from i2b2.org, or clone the Github repo.
- Navigate to the data package, under Release_1-8/Upgrade/Metadata and edit the db.properties file as needed.
- If your i2b2 is below version 1.7.12a, run the ant upgrade script . This will create to create the totalnum and totalnum_report tables. See the upgrade guide for complete information.
e.g., ant -f data_build.xml upgrade_tables_release_1-7-12a - In the Release_1-7/NewInstall/Metadata/ run the Run the ant script target to create the stored procedures.
ant -f data_build.xml createxml rebuild_metadata_procedures_release_1-7 - Set privileges: If using multiple schemas, the stored procedure should be run from the metadata schema. Make sure the stored procedure can read the tables in the crcdata schema (observation_fact, visit_dimension, patient_dimension) and can both read an update ontology tables in the metadata schema (including table_access).
The code for the totalnum stored procedures can be found here.
Fast Totalnum
i2b2 1.8 introduces a version that is 5-10x faster. This faster version is presently only available for MSSQL and has only been extensively tested with the ACT ontology. 1.8.1 and later x versions will improve on this faster version. These replace the pat_count_dimensions
and run_all_counts
stored procedures.
...
If using multiple fact tables, the recommended approach is to create a fact table view as the union of all your fact tables. (This is essentially going back to a single fact table, but it is only used for totalnum counting. This is needed to correctly count patients that mention multiple fact tables within a hierarchy.)
e.g., create view observation_fact_view as select * from CONDITION_VIEW union all select * from drug_view
- If running the counting script in SQL Server, add the wildcard flag, to ignore multifact references in the ontology: e.g. exec RunTotalnum 'observation_fact_view','dbo','@','Y'
This is automatically accounted for in the other database platforms. Note this approach does not work if you have conflicting concept_cds across fact tables.
- If running the counting script in SQL Server, add the wildcard flag, to ignore multifact references in the ontology: e.g. exec RunTotalnum 'observation_fact_view','dbo','@','Y'
Execution
See database-specific instructions below. After running the scripts, results are placed in: c_totalnum column of all ontology tables, the totalnum table (keeps a historical record), and the totalnum_report table (most recent run, obfuscated). These total counts will also be visible in the ontology browser in the web client.
MSSQL Version
By Mike Mendis and Jeff Klann, PhD based on code by Griffin Weber, MD, PhD
Run with:
exec RunTotalnum or exec RunTotalnum 'observation_fact','dbo','@'
The optional parameters are:
- Observation table name (for multi-fact-table setups)
- Schema name
- A single ontology table name (specify to to run on a single ontology table - otherwise (or if '@' is specified) runs on all tables in table_access)
- A wildcard flag that will ignore multifact references in the ontology if 'Y'. (See below for the use case.)
Note that visit and patient dimension will only be counted in conjunction with the default (observation_fact) tablename!
To use with multi-fact-table setups:
Option 1) If you have at most one fact table per ontology, run this once with each fact table specified!
e.g., to use on a fact table called derived_fact with just the act_covid ontology: exec RunTotalnum 'derived_fact','dbo','act_covid'
Option 2) Create a fact table view as the union of all your fact tables. (This is essentially going back to a single fact table, but it is only used
for totalnum counting. This is needed to correctly count patients that mention multiple fact tables within a hierarchy.)
e.g.,
Example 1: Counting using OMOP tables
create view observation_fact_view as
select * from CONDITION_VIEW
union all
select * from drug_view
And then run the totalnum counter with the wildcard flag, to ignore multifact references in the ontology, e.g.,
exec RunTotalnum 'observation_fact_view','dbo','@','Y'
Example 2: Counting using a derived fact table and the regular fact table, using a single ontology
create view observation_fact_view as
select * from observation_fact
union all
select * from derived_fact
Run the totalnum counter with the wildcard flag, to ignore multifact references in the ontology, and specify an ontology table, e.g.,
exec RunTotalnum 'observation_fact_view','dbo','act_covid_v4','Y'
Note this approach does not work if you have conflicting concept_cds across fact tables.
Oracle Version
By Mike Mendis, based on SQL Server code by Griffin Weber, MD, PhD
Performance improvements by Jeff Green and Jeff Klann, PhD 03-20
Run the procedure like this (but with your schema name instead of i2b2demodata):
begin
runtotalnum('observation_fact','i2b2demodata');
end;
You can optionally include a table named if you only want to count one ontology table (this IS case sensitive):
begin
runtotalnum('observation_fact','i2b2demodata','I2B2');
end;
Note: If you get the error as: ERROR at line 1: ORA-01031: insufficient privilege, then run the command:
grant create table to (DB USER)
Postgres Version
Original PostgreSQL code by Dan Vianello, Center for Biomedical Informatics, Washington University in St. Louis
2019 - Modified for i2b2 1.7.12 release by Mike Mendis, Partners Healthcare
2020 - Updated to support reporting and single-table runs by Jeff Klann, Massachusetts General Hospital
Usage example:
select runtotalnum('observation_fact','public')
- Replace 'public' by the schema name for the fact table.
- If using a schema other than public for metadata, you might need to run "set search_path to 'i2b2metadata','public' " first
- You can optionally specify a single table name, to count using only one ontology table. This is case sensitive.
Running using ANT
Run the ant command to execute the data_build.xml file with below specified target
- POSTGRESQL : ant -f data_build.xml db_metadata_run_total_count_postgresql
- ORACLE : ant -f data_build.xml db_metadata_run_total_count_oracle
- SQL SERVER : ant -f data_build.xml db_metadata_run_total_count_sqlserver
Some additional notes on running Postgres
...