Qlik Sense Script development and standardization atempt
Focus: Mature organizations will have many large models in play but for the sake of this challenge, lets say 20 large focused models consisting of a minimum of 12 tables in each model. The new challenge is to combine multiple models possibly 3 into one super App. This request will be one of many looking forward. So my thought is to store those focused model into a text (qvs) file. Then those complete models can be loaded via script to control delivery of tested models. I included a test file below.
My question is:
Is there anyway to Read, Run & load proven script as in the example above from either a file or $(Must_Include) statement?
If you really want to go further with this topic you will probably need to unify all relevant data-models. This might be done by developing everything as a star-scheme data-model - means a single fact-table and n dimension-tables. Further the fact-tables should have a real stream-data-structure - this means there exists only a single value-field for each kind of measure regardless if sales or budget or amount or anything else. To differentiate between them you add a KPI field which contained the name. A similar logic might be used to associate some of the dimensions to it by putting everything like a customer-id or store-id in an ID field and applying a further field ID-type.
In reality this won't do what do you have described but the effect of it may come quite near. Means the loading of n different data-areas from appropriate prepared qvd's could be done with a single load-routine - whereby routine means routines with with n nested loops over folder and files and reading various information from the folder- and file-names like the YYYYMM periods to specify exactly within the load-call which data should be loaded.
If you really want to go further with this topic you will probably need to unify all relevant data-models. This might be done by developing everything as a star-scheme data-model - means a single fact-table and n dimension-tables. Further the fact-tables should have a real stream-data-structure - this means there exists only a single value-field for each kind of measure regardless if sales or budget or amount or anything else. To differentiate between them you add a KPI field which contained the name. A similar logic might be used to associate some of the dimensions to it by putting everything like a customer-id or store-id in an ID field and applying a further field ID-type.
In reality this won't do what do you have described but the effect of it may come quite near. Means the loading of n different data-areas from appropriate prepared qvd's could be done with a single load-routine - whereby routine means routines with with n nested loops over folder and files and reading various information from the folder- and file-names like the YYYYMM periods to specify exactly within the load-call which data should be loaded.
