27 June 2021 Release Notes¶
Version 1.1.221. Note that version 1.1.220 was deployed to DataKitchen testing environments only.
Feature Highlights¶
File-Based Connectors Support Folder Wildcards¶
When configuring a wildcard mapping to file-based data sources and data sinks, users can now employ glob syntax (*) to retrieve files from or send files to multiple folders in a path. Both the Folder Path (for sources) and Target Folder Path (for sinks) fields support a wildcard in the final segment of a path.
File-based connectors include Azure Blob Storage, Azure Data Lake Storage Gen2 (ADLS2), FTP, Google Cloud Storage (GCS), Amazon S3, and SFTP. See File-Based Properties and File-Based Wildcards for more information. [DEV-8258, DEV-7993, DEV-8171]
Note
Wildcard Match Includes Subfolders: At this time, a folder wildcard directs the system to look for files in the designated folder and in any subfolders of that folder.
Wildcard mapping configured to retrieve multiple CSV files from multiple folders on an SFTP server.
Recipe README Gets a Wiki Editor¶
The new Wiki tab on the Recipes page now offers an editor for modifying README.md files directly in the user interface!
The editor includes a standard toolbar for text styling, inserts, and previewing. See Use the Recipe README for more information. [DEV-7790]
Users can edit recipe READMEs in the Wiki tab.
Other Fixes & Enhancements¶
Merge Prompts Prevented if No Changes: In a scenario where changes were made in a child kitchen after parent and child kitchens were completely synched with merges, the platform was prompting users to merge down again prior to merging the new changes up. The merge down would have zero changes in this case, so the warnings were meaningless.
This problem stemmed from the way Git recognizes merges as changes and considers a parent kitchen branch to be one step ahead when a child merges into it. This issue has been resolved by synching parent and child to the same point in history in a "fast-forward" merge, so that the system will not assume a change is available for merging down. [DEV-7150]
Node Status Displays Properly for All Node Types: Earlier this year, Automation introduced a new node status in the Order Run Details graph. A node displays the yellow Completed + Warnings status if it completes its execution but has failed tests that were configured with a warning failure mode. Recently, users reported that some nodes were not displaying this status when they should have. The problem stemmed from the system not checking all possible code locations for test definitions. This issue has been resolved, and all nodes with warning-mode test failures now indicate the correct status. [DEV-8378]
Nodes with warning-mode test failures display with the Completed + Warnings status.
User's Selection Persists for Logs: Users routinely change the Order Run Logs dialog to show more items per page. This release updates the display options to 100 (default), 200, 500, 1000, and 2000 items per page. A user's selection is now stored in the browser, so it persists across user sessions until the user clears the cache or opens a different browser. [DEV-8259]
Logs Retain Page Context on Refresh: The Order Run Logs dialog was refreshing at scaling intervals when an order run was executing, and on refresh, the dialog would reset to display the first page of the logs. Users who navigated past page one found they were sent back every time. This issue has been resolved; the refresh interval was updated to a standard five seconds, and the user's page context is retained when the logs are refreshed. [DEV-8508, DEV-8511]
Chrome Auto-Complete Disabled for Input Fields: Google Chrome's auto-complete feature was obstructing some platform input fields. This issue has been resolved, and the Chrome feature is disabled for platform fields that offer drop-down lists or type-ahead, auto-complete functionality. [DEV-8283]