Helpful Resources
- For installation and configuration, please refer to the WebAPI Installation Guide.
- For an overview of ATLAS please refer to the Atlas Video Tutorials.
Important Upgrade Notes
- You must configure one of your vocabulary daimon's to have a priority of 1 or higher as part of this release.
- If you have entities (i.e. cohort definitions, concept sets, etc) with names that are 250 or greater characters in length, you may experience an issue whereby the migration script hangs. This is a known limitation described in #1105 (comment). We'd suggest you use SqlRender to translate the query below and run it against your WebAPI database (specifying the
ohdsiSchema
for your environment) to see if there are any entities with longer names that should be corrected before upgrading:
{DEFAULT @max_length = 250}
select 'concept_set' table_name , concept_set_name entity_name from @ohdsiSchema.concept_set group by concept_set_name having COUNT(*) > 1 AND LEN(concept_set_name) >= @max_length
union
select 'cohort_definition' table_name , name entity_name from @ohdsiSchema.cohort_definition group by name having COUNT(*) > 1 AND LEN(name) >= @max_length
union
select 'cohort_characterization' table_name , name entity_name from @ohdsiSchema.cohort_characterization group by name having COUNT(*) > 1 AND LEN(name) >= @max_length
union
select 'fe_analysis' table_name , name entity_name from @ohdsiSchema.fe_analysis group by name having COUNT(*) > 1 AND LEN(name) >= @max_length
union
select 'pathway_analysis' table_name , name entity_name from @ohdsiSchema.pathway_analysis group by name having COUNT(*) > 1 AND LEN(name) >= @max_length
union
select 'ir_analysis' table_name , name entity_name from @ohdsiSchema.ir_analysis group by name having COUNT(*) > 1 AND LEN(name) >= @max_length
union
select 'estimation' table_name , name entity_name from @ohdsiSchema.estimation group by name having COUNT(*) > 1 AND LEN(name) >= @max_length
union
select 'prediction' table_name , name entity_name from @ohdsiSchema.prediction group by name having COUNT(*) > 1 AND LEN(name) >= @max_length
;
Fixes
- Data Sources
- Cohort Definition
- Add more descriptive error if cohort fails to generate (#972)
- Remove generating jobs for deleted cohort definitions (#916)
- Split reporting queries into separate files (#385)
- Fixes retrieval of Heracles treemap drill-down reports (#1126)
- Fixed error in procedures by type report SQL (#1151)
- Fixes condition & drug era treemap drilldowns in cohort reporting (#933)
- Revise concept path in cohort results queries (#1195)
- Cohort Characterization
- Moved characterization code into external skeleton package (#975)
- Fixed approach for updating feature analysis (#785)
- Fixed generation problems on BigQuery (#1120)
- Replace design hash value from results with human-readable display (#1110)
- Save analysis information before completion (#1174)
- Fix generation issues (#1188)
- Cohort Pathways
- Incidence Rates
- Display an error if you do not have permissions to create an Incidence Rate analysis (#948)
- Fixes permission checks when generating IR on a data source (#1130, #1123, #1740)
- Use full cohort definitions for import/export (#1171)
- Fixes permissions for imported IR analysis (#1183)
- Fixes export to csv (#1879)
- Estimation/Prediction
- Security
- Fixed role permissions (#989)
- System
- Ensure all entity names are unique (#1307, #1338, #1146, #1192)
- Optimize generation retrieval for characterization & pathways (#828)
- Companion updates for the wiki (#999)
- Removed non-OHDSI references from pom.xml (#1051)
- Fixed Oracle DDL results schema (#1037)
- Add a caching mechanism to the stack (#1070)
- Utilize escape sequences with LIKE searches in Hibernate (#1153)
- Fixes for Oracle migrations (#1159)
- Improved debugging capabilities for Hikari pool of custom auth DB (#1185)
- Changes log level for execution engine (#1200)
Here is the full list of fixes
Migration Notes
Use SqlRender to translate the following SQL for your specific environment and run on all CDM results schemas:
IF OBJECT_ID('@results_schema.pathway_analysis_codes', 'U') IS NULL
CREATE TABLE @results_schema.pathway_analysis_codes
(
pathway_analysis_generation_id BIGINT NOT NULL,
code BIGINT NOT NULL,
name VARCHAR(2000) NOT NULL,
is_combo int NOT NULL
);
IF OBJECT_ID('@results_schema.pathway_analysis_paths', 'U') IS NULL
CREATE TABLE @results_schema.pathway_analysis_paths
(
pathway_analysis_generation_id BIGINT NOT NULL,
target_cohort_id INTEGER NOT NULL,
step_1 BIGINT,
step_2 BIGINT,
step_3 BIGINT,
step_4 BIGINT,
step_5 BIGINT,
step_6 BIGINT,
step_7 BIGINT,
step_8 BIGINT,
step_9 BIGINT,
step_10 BIGINT,
count_value BIGINT NOT NULL
);
Acknowledgements
Thanks to @pavgra, @ssuvorov-fls, @aklochkova, @AntonStepanof, @chrisknoll, @wivern, @fdefalco, @AnthonyMolinaro for their contributions and reviews as part of this release.