How to change a training set from external to internal

Issue

By default, in a new Classification or Extraction Group the training sets will be stored internally (i.e. in the database), however some customers then choose to use an external folder as the training set instead (e.g. C:\ExternalTrainingSetFolder).

Sometimes this is done unintentionally and there is no obvious way of setting back to an internal training set.

Therefore in order to revert from an external training set back to an internal training set again, please follow the below steps:

Solution

1. Open the Classification Group/Extraction Group (CG/EG) in Transformation Designer (TD)
2. In 7.7+ you can open the project folder by clicking on Project and then Project path. See image below.                                                                                        
   • Otherwise to find the project folder follow the below steps:
     • Open the current user’s temp folder by navigating to “%temp%” in Windows Explorer
     • The currently open project will be in a subfolder with a temporary name such as “rztz4ghl.4hf”. It will most likely be the most recently modified folder. For example my temp folder looks like C:\Users\Administrator\AppData\Local\Temp\2\dztz4ghl.4hf . You can also hover over the training set in TD to see the full temp path. Make sure to hover over the partially shown path to see it in full (as below): 
        
     • 
3. Open the ClassificationTraining or ExtractionTraining folder within the project folder.
4. Copy the contents of the current external training set (i.e. the set residing on the external folder) into this ClassificationTraining or ExtractionTraining folder.
5. After the copy completes, open the ClassificationTraining or ExtractionTraining folder (in the user's 'temp' directory) as a test set in Transformation Designer.
6. Right click on the newly opened test set and click “Use as Classification Training Set” or “Use as Extraction Training Set”, click OK if prompted to confirm.
7. Train the project (Process > Train > Classification/Extraction)
8. Release the project, which will now include the training set when saving back to the database.

Level of Complexity 

Moderate

Applies to  

References