Changes in version 4.2.0 (2026-08-25) New Features Two new options are available on the IncidenceDesign: - firstAtRisk: limits TARs to the first continuous TAR period - firstPostOutcome: Only uses outcomes that occur within clean window of TAR or later. Changes in version 4.1.1 Bug Fixes Fixed issue to exclude persons who TARs push them out of cohort observation period. Changes in version 4.1.0 =========== New Features You can now specify age breaks as a list via the ageBreakList paramater in StrataSettings. Bug Fixes Fixed issue where excluded persons was not calculated correctly. Adjusted validation of name properties to be 'string'. Changes in version 4.0.0 =========== Major Changes The results DDL has been split into 5 tables: - incidence_summary - target_def - outcome_def - tar_def - age_group_def - subgroup_def executeAnalysis() will return a list of dataframes with the following fields: - incidenceSummary - targetDef - outcomeDef - tarDef - ageGroupDef - subgroupDef These dataframes follow the same structure as the corresponding table. Changes in version 3.3.1 =========== Bug Fixes - Fixed summing large number (bigint) of days by casting to bigint. Changes in version 3.3.0 =========== Major Changes - Performance release: improved IR Calculation SQL. Changes in version 3.2.0 =========== Major Changes - Performance release: improved era-building SQL to improve performance. Changes in version 3.1.5 =========== Bug Fixes - Fixed strata settings when index-year strata was applied when gender was applied. Changes in version 3.1.4 =========== Bug Fixes - Fixed non-deterministic query when finding end dates of eras. Changes in version 3.1.3 =========== Bug Fixes 1. Fixed Bug: Encode input strings to prevent sql injection (#29) 2. Fixed Bug: Truncate target and outcome names to 255 characters. Changes in version 3.1.2 =========== Bug Fixes 1. Fixed bug: Exlusion time was not being calculated when clean window = 0. Changes in version 3.1.1 =========== Bug Fixes 1. Fixed SQL error when all strata settings were set to F. Changes in version 3.1.0 =========== Major Changes 1. Implemented StudyWindow setting: TAR will be censored at the study window end, and excluded if TAR did not start during the study window. Changes in version 3.0.1 =========== Bug Fixes 1. Fixed issue when age stratify is false. Changes in version 3.0.0 =========== Major Changes This release introduces R6 classes to encapsulate the elements of an Incidence Design. See documentation and vignettes for information on use of these new classes. Note: the createXXX() methods which formerly returned a jsonlite-compliant list now return R6 classes. Therefore, the release was incremented by a major version. Changes in version 2.0.0 =========== Major Changes This release introduces new features and an altered results schema table, requiring a major version increase. Features: 1. New 'strataSettings' option in the CohortIncidence design that can stratify by age, gender and start year. 2. New 'useTempTable' option to do everything in temp tables, with an additional 'cleanup' function to remove tables post-execution. 3. incidence_summary table had some columns renamed for clarity 4. Results schema has columns and column values changed (example: TAR_START_INDEX 0=end, 1=start changed to TAR_START_WITH 'start' = start and 'end' = end) Changes in version 1.0.1 =========== Bugfixes: 1. Fixed missing date padding by adding +1 to sum(start-end). 2. Fixed lean window dates by setting timespan to start+1 .. end+cleanWindow Changes in version 1.0.0 =========== Initial Release Including: 1. Java implementation 2. Standard object model for defining CohortIncidence design. 3. Test cases