> For the complete documentation index, see [llms.txt](https://docs.collibra.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.collibra.com/data-quality-self-hosted/release-notes/data-quality-observability-self-hosted-release-archive/release-2021-12.md).

# Release 2021.12

#### \*Note to Standalone Collibra DQ Customer Upgrades\*: We have upgraded to Log4J 2.17, please refer to Standalone Upgrade for additional steps

#### Enhancements

* Rules
  * Semantic and data concept management: Run Discovery feature
    * Run Discovery feature can be accessed from Catalog by selecting 'Data Concept' option from Actions or clicking the 'Run Discovery' button on the Rules tab of the DQ Job page. This will run a DQ Scan to detect for the semantics assigned to the selected data concept
    * Algorithm now selects best match if column matches 2 or more data classes based on % match and name distance
  * \*Tech Preview\* \[TP] Configurable rule break preview limit
    * Global default is 6 max rows per rule
    * Any change from 6 must be specified with previewLimit (API /v2/createrule) or in the Preview Limit field (UI)
    * Maximum of 50 from UI
  * Introducing additional Stat Rules including minPrecision, maxPrecision, minScale, maxScale
* Behavior
  * Min and max value checks are now triggered for all numeric columns when selected, even if column contains zeroes in lookback period
  * AR column view graph now shows theMean value for current day (runId). No re-run of DQ Job is necessary. The displayed Mean makes it clear that the % change is the % change from the mean, not runId - 1 day.
  * Findings in behaviors that were directly correlated to a row count shift as the root cause have been optimized, such that a major deviation in row count will no longer down-score related fields in the dataset to reduce noise
* Catalog
  * Catalog now features intelligent ranking based on Recency, Most Scanned, User
* Outliers
  * Dynamic minimum history allows for gaps in dates when establishing lookback period, which is established by history with row count > x (specified by user)
  * Fixed issue where outlier data preview graphics were not displayed
  * Fixed issue where outlier results did not honor the initial scope where clause, in particular for Remote Files (S3)
* Connections
  * BigQuery: Enhanced support for cataloging host name
* Pulse View
  * Pulse view can filter Connections and Users
  * Pulse view can serve as proxy verification on whether scheduled jobs were successfully completed
* Profile
  * Viewable precision and scale statistics for double, float, and decimal data types
* Shapes
  * Fixed issue where data shape preview not available when same shape is detected on the same row for different columns
* Files
  * \*Tech Preview\* \[TP] Users have ability to assign an agent when using temp file and local file Explorer paths
    * Known limitation: -master must be freeform appended to the agent or to each job
  * Support for multicharacter delimiters
  * Improved delimiter support to distinguish string commas versus actual CSV commas to align data to respective columns
* Agent
  * Fixed issue where certain completed jobs could not be re-run on the DQ Jobs page. In other words, NO\_AGENT was the only available option in the Agent dropdown. Now, users can select valid agents in the dropdown and this will persist for future scheduled jobs
* Schedule
  * Implemented validation to enforce user to choose days when picking schedule to avoid Java error messages
* Explorer
  * Fixed issue where '&' was not properly supported when adding additional parameters
* API
  * JSESSIONID session time is configurable
  * Bearer token and JSESSIONID authentication paths are properly forked
* Pattern
  * Patterns activity now shows Count (number of times the current dataset has the Pattern breaks). This Count is interpreted the same way as Outlier activity Count