@Dmitrij Mamajev,
Blue Prism offers a set of Utility Programs called, 'Decipher Training Data Processes'. Within that Release, I would recommend 'Decipher - Extract Training For Document Type'. At this point, the process is a bit tedious, depending on how many Decipher automations/projects you have in your PROD environment. My steps runs generally as follows:
1. Take a full extract of the Training Data from your Environment
2. Run the 'Decipher - Extract Training For Document Type' process for each active Document Type. This should leave you with individual Training Data files.
3. Delete ALL Training Data from your Environment
4. Load the individual Training Data files as required.
Obviously, I hope, test this in your lower environments before proceeding with PROD. But, you will have the full extract should things go awry.
(Link to the documentation and download link: https://bpdocs.blueprism.com/decipher-2-1/en-us/Guides/processes/training-data-process.htm).
We have also instituted policies to completely clear the lower environments of Decipher of all Training Data before developing a new process or remediating a break/fix (existing projects start with Training Data extracted from PROD through the steps above). This is meant to ensure that only related Training Data is promoted to Production.
My hope is that the changes that have been rumored for Decipher 2.3 will make the above obsolete.
Let me know what you think,and good luck,
Red
------------------------------
Robert "Red" Stephens
Application Developer, RPA
Sutter Health
Sacramento, CA
------------------------------
Robert "Red" Stephens
Application Developer, RPA
Sutter Health
Sacramento, CA
