Unclear workflow

[schuemie]

Maybe its me, but I found the DQD package workflow a bit unclear. I ran executeDqChecks(), and this secretly created a folder with a JSON file that had an auto-generated name. It also produced a lot of output, so apparently I was supposed to capture the output of this function (although I'm not sure what I should have done with the output). I then ran viewDqDashboard() but this required the path to a JSON, which after some puzzling I realized is the auto-generated name of the JSON in the secretly created folder.

In general, simply writing to somewhere on the users file system is discouraged (CRAN would reject the package for this practice), so I recommend making outputFolder a required argument of executeDqChecks() (meaning: don't set it to a default value, and throw an error if the user doesn't provide it).

I'm not sure why the name of the JSON is auto-generated. That means the user will need to go to the folder to figure out the name. Why not use a fixed name (like 'dqd.json')? Otherwise, make sure the executeDqChecks() function lets the user know the name of the JSON file.

I'm not convinced executeDqChecks() must return a large complex object in addition to writing everything to the JSON, But if you are, then may I recommend returning the object invisibly, so as not to clutter the console?

It would be nice if we could simply point viewDqDashboard() to the same output folder used in executeDqChecks(), and it would automatically detect the JSON file. That would probably require a fixed name for the JSON though.