New Index Cluster Type

[JÞB]

I wrote code similar to the example the other day and thought to myself.  "Jeff, This is Stupid! Why am I unbundeling the row and column to feed into index array?"

If only we had a new cluster of type "index" that contains 1-n I32 scalars we coud index arrays by row, column, page, book. ad nausuem. with a simple index cluster wire.  For IDEAL functionality a -1 element would select none of that dimention and -2 would select ALL of that dimention to slice rows, columns etc...

It should be fairly easy to integrate with a RCM on the index terminal of array funtions as Create>>Index cluster. and BANG a constant cluster adapted to the current dimentionallity of the array is dropped on the BD.  Right Click the index cluster to Add dimention or Remove dimention too.  Array indexing made easy!    

[JÞB]

OK I see a problem with the -2 part- it would make the output dimetionality indeterminate at compile time so scratch that (unless its possible somehow.)

Another neat feature would be to allow the elements to grab a reference to a iteration terminal like so:

Wait, you say, that's auto-indexings job!  ABSOLUTELY! but with higher dimentions this allows some unique behavior.  and no need to transpose arrays if you want to index by Col,Row.  Variable + i is possible too.

[smercurio_fc]

Peanut butter + chocolate = awesome combination.  Ah, how I miss the days of being able to eat chocolate.

Clusters + Index Array = not so much. I'm thinking about the use cases for this and it does not seem to me to be that big. That, coupled with the confusion it would cause for new users (they already have enough problems trying to get the fact that you can resize Index Array, or how to select only columns or rows), makes me think we should not cross this line.

[Darren]

Any idea that has received less than 3 kudos within 3 years after posting will be automatically declined.