IS REF any good (cheeky question)

This is a genuine question to see what real life production experience is like. The real question is in real life do you modify REF to your own standards? Please forgive any stupidity :wink:

What’s driving the question? Working on the Client Hash training example and having lots of juicy problems which are fun to solve (I am trying not to cheat and just ask on the forums). I think, but could be wrong, the walk through is out of date.
REF looks great when using queues, but the specific example in this exercise seems wierd.

  • I get a datatable
  • Then I convert it to a an array of datarows seemingly for the sole reason that I can retrieve items from it using an auto generated index which is incremented later on. Seems a bit wierd. Surely I would just iterate over the data table and generate a key as a reference?
  • The other thing that confuses me is it is kinda missing a state in it’s current form. All my data prep and gathering of an overall data set is in initialise applications. Wouldn’t it be better off having a seperate state of data prep. In real life there is usually a ton of difficulty with databases, apis and other systems PLUS ensuring all data is correct and formatted (validation and enhancement perhaps) which would logically be another state. Yes you can do that in each transaction item, but it seems inefficient.

I would apreciate any views. Or how you are all modifying it for production use if you are.

Check it out…
This might help you…