Annotatio
This is a new release with a lot of new features, bug fixes and improvements.
* **Case Definition** in Analysis Controls. Now controls are more about cases than errors. Users can define a list of cases at the control level and map each data record to a specific case. In turn, each case has its own textual representation shown in the result tables. On top of that, a case can have one of the following types: Normal, Info, Error, Warning, Incident, Discrepancy. Cases other than Normal are automatically saved into the result table as control errors.
* **Error Definition** now can be done by **SQL-like** expressions instead of using JSON configuration, which is kept anyway for API or possible Service integrations.
* Datasources can be **filtered** now with **SQL-like** expressions inside the control configuration table to avoid extra usage of views instead of ordinary tables.
* Now users can configure a control **Preparation SQL** which performs before the control. This allows to trigger some things before the control or even execute PL/SQL ETL procedures to prepare control datasources.
* Now users can configure a control **Prerequisite SQL** to check some conditions before the control. If this check is not passed, then control is not resumed.
* Redesigned the control spawn method, which now not only launches but also **handles** the control in a separate **background** threads and, as a result, releases the main thread or user API request.
* Control results now can be deleted each time the control executes.
* Control result tables now can be dropped and re-created each time the control executes.
* Faster deletion of control results that must not be kept more than a day.
* Now PL/SQL **HOOK** procedure become **POSTRUN** hook combined with a new **PRERUN** hook. Thanks to krokce. The former `need_hook` parameter from the control configuration table became a global value defining the status of those two.
* Better control configuration audit. Now users can see when and who has updated the control. Also, OS logins now used instead of DB usernames.
* Fixed control cleanup functionality. Thanks to krokce.
* Fixed database connection overflow and timeout issue when exceptions in threads occur.
* Added new parameters to the configuration file for better database connection setup.
* Extended the max control name length to 45 characters, modified other parameter types and length limits for better user experience and Oracle 12.2+ compatibility.
Important Changes
This release brings some notable database schema and configuration changes described below.
Database
* New fields in `RAPO_CONFIG`: `source_filter`, `source_filter_a`, `source_filter_b`, `case_config`, `result_config`, `with_deletion`, `with_drop`, `need_prerun_hook`, `need_postrun_hook`, `preparation_sql`, `prerequisite_sql`, `updated_by`, `updated_date`.
* Some of the VARCHARs in `RAPO_CONFIG` become CLOBs, some others have increased in length. See the updated DB schema file.
* New fields in `RAPO_LOG`: `prerequisite_value`, `text_message`.
* New dictionary with Case types description `RAPO_REF_CASES`.
* New triggers to catch `RAPO_CONFIG` changes.
* Added `RAPO_PRERUN_CONTROL_HOOK` function. Renamed `RAPO_CONTROL_HOOK` to `RAPO_POSTRUN_CONTROL_HOOK`.
Configuration
* Added `DATABASE` parameters: `driver_name`, `client_path`, `max_identifier_length`, `max_overflow`, `pool_pre_ping`, `pool_size`, `pool_recycle`, `pool_timeout`.
* Changed `DATABASE` parameters: `username` (former `user`), `vendor_name` (`vendor`), `service_name` (`service`).
See commits of this release [here](https://github.com/t3eHawk/rapo/compare/v0.3.3...v0.4.0).