Subset Loads
Sometimes it is not necessary to load the whole logic when making minor adjustments. It is possible to do a Subset Load or also called Partial Load instead of loading everything over and over again. This reduces the computing time and is therefore more efficient. This is especially useful when the data model is very large and you only want to adjust, for example, individual MasterItem translations in the FrontEnd or have made adjustments to the RCA or Conformance Checking parameters. In order to shorten the loading times for such minor adjustments, only the adjusted logic sections can be called up.
To do a Partial Load we will need a Subset Load Template (corresponds to the QVS logic) and the variable SET mpmXvAppConfigAlias
, which can be set, for example, in the mpmX TemplateApp TaskApp [[Version]].qvd or in the mpmX TemplateApp [[Version]].qvd.
The following three Subset Load templates are available. They are located under mpmX_scripts/[[version]]/customizing/_templates:
- custom_logic_onlyMining.qvs
- custom_logic_onlyFrontEnd.qvs
- custom_logic_onlyRCAandCC.qvs
In another window open the DataLoad Editor inside of the mpmX App, where the SET mpmXvAppConfigAlias
variable can be defined.
Subset Loads - onlyMining​
The onlyMining-Load generates the core mpmX model by mining the provided EventLog table. It generates the process log, identifies variants, and prepares case-specific information, such as timestamps and calendar data.
-
Copy the custom_logic_onlyMining.qvs file and paste it into your process folder which in this example is mpmX_script/[[Version]]/customizing/ExampleProcess.
-
Adapt the
mpmX_execute
routine to your EventLog/ResourceLog datasets and decide what else should or should not be calculated (e.g. RealUser, ReworkEvent).
If LET mpmXvStoreAnalyticProcessModel
is set to 1, it will store the model locally. The model will be saved under mpmX_data/process_analytic_models/[mvProcessDescription]. If there is nothing saved, this is because no routine was executed yet.
-
Go back to the DataLoad Editor and change the variable
SET mpmXvAppConfigAlias ='onlyMining'
. It will then only access the .qvs file which hasonlyMining
in its name instead of the custom logic from the whole model. -
Still in the DataLoad Editor click on "Load" in the top right corner and it will only do the mining.
Since we have LET mpmXStoreAnalyticProcessModel = 1
set to 1, we can now see under mpmX_data/process_analytic_models/ExampleProcess that the data model got stored.
This is important when doing several partial loads, which will be explained down below.
Subset Loads - onlyFrontEnd​
The onlyFrontEnd-Load loads the mined mpmX core model from the QVD layer if it is not already in memory. It then prepares UI translations for active languages and loads master item definitions. The model content is then translated, including the creation of additional fields or tables, and search index generation is configured based on include/exclude rules in the mpmX_repository.
Similar to the onlyMining we can only load the Front End by using the template custom_logic_onlyFrontEnd.qvs.
-
Copy the template custom_logic_onlyFrontEnd.qvs into your ExampleProcess Folder.
noteYou do not have to change anything in the script logic within the custom_logic_onlyFrontEnd.qvs. However, it is important that the onlyMining scenario has been run at least once before executing the onlyFrontEnd scenario and that the data model has been saved before (the parameter
mpmXvStoreAnalyticProcessModel
must be set to the value 1 for this - see again Step-by-Step Instructions from Only Mining). -
Back in the application we need to change the alias variable to
SET mpmXvAppConfigAlias = 'onlyFrontEnd'
-
Click on Load again. This will basically just refresh the Front End. This can be useful if you want to make changes to the MasterItems or Translations without having to recalculate the whole mining process.
Subset Loads - onlyRCAandCC​
This subset load is primarily used to reload changes made to the Root Cause Analysis and Conformance Checking parameterization. Accordingly, changes should have been made in the mpmX repository in advance, e.g. to the HappyPathTable, RootCauseAnalysisDimensions or RootCauseAnalysisFlags inline tables.
-
Just like we did before, copy the custom_logic_onlyRCAandCC.qvs from the Templates-Section and paste it into your ExampleProcess [or your own Process] Section.
noteYou do not have to change anything in the script logic within the custom_logic_onlyRCAandCC.qvs. However, it is important that the onlyMining scenario has been run at least once before executing the onlyRCAandCC scenario and that the data model has been saved before (the parameter
mpmXvStoreAnalyticProcessModel
must be set to the value 1 for this - see again Step-by-Step Instructions from Only Mining. -
Go back into the DataLoad Editor in your App. Set the Alias to
SET mpmXvAppConfigAlias ='onlyRCAandCC'
. -
Load this logic. This will update only the corresponding tables in the mpmX data model, which are required for the correct analysis and representation of the RCA and Conformance Checking.