Suggestion regarding the topic of PolarhistogramFS fuction

[cubansupernova]

In the polar histogram section, it may be helpful to add a brief debugging note reminding readers to check that the grouping vector does not have a higher number of elements as the data (e.g. height(T) >= numel(zone)), which is not a problem when the data is organized but can lead to mistakes if slicing and putting back together is needed. Off-by-one mismatches can otherwise lead to error messages that aren't very informative.

PolarHistogram error

PolarHistogramFS error

[MarcoRianiUNIPR]

@cubansupernova I understand your point and thank you for your suggestion.

Sometimes in the code when you try to be very general it is difficult to cope with everything.

In line 209 of polarhistogramFS we try to investigate whether groupvariable is external or internal to the table

Our check is that if the length of the supplied variable is the same of the number of rows of the table then we understand it is an additional variable (see line 211).

Other times grouping variable is passed as a char and this char has a length greater than 1 but in this case this char refers to the name of a variable inside the table.

I have to think of how to cope with all these possibilities. Any suggestion is welcome

[BoschiGiacomo]

Good evening, i have quickly reviewed the code of the function, as i'm busy preparing my next exam right now, so i don't have the time to draft a fully fleshed out solution, unfortunately.

The fastest fix i thought of is to nest another series of if inside the argument validation block, first checking if the grouping variable is or not inside the table, and then letting the new series of ifs handle the errors. This would allow us to handle exceptions with more specificity, without relying on that last "else" block that covers all the exceptions. However i consider this to be a "band-aid" solution and not one that i would be happy with, as nesting another set of if statements inside that block would make the code messier and less readable.

I'm going to draft a better alternative suggestion in my spare time and post it as soon as i can test it.