There is often a need to control the order of running of data functions, or to only run a data function when inputs have changed. Another common use case is to automate these data functions through Automation services.
Typically this can be controlled by setting data functions to refresh automatically, so that they run without interaction and react to changing data from other data functions or inputs. However, when using refresh automatically, all data functions automatically run when a DXP opens regardless of any of the inputs have changed or if caching is turned on. This makes chaining and automating data functions difficult.
An example is a data preparation data function that passes its results to a data science data function. Currently, on opening the DXP both data functions run instead of the data science data function running once the data preparation function completes.
Therefore, it would be very useful and help with hyperconverged analytics if it was possible to either set whether a data function runs on opening of the DXP, or another option may be to allow caching to persist between sessions. Being able to prevent a data function from running on initial option is likely better for the scenarios described here.