Derive poisson probability values as "grouped by"

[awesome_opossum]

What I'm trying to do is derive the probability of values on the basis of a Poisson function, but by variables in long form, such as would be the group by function in proc sql.  The data step has the pdf() function, which can return the Poisson probabilities, but has no group by function.  

I'm aware it's possible to do a 'group by' on operators using first. and last. in data step, but surely this wouldn't work for a function like pdf(). 

There may be procedures that can do this, but I'm unaware of what they are. 

Does anyone have any recommendations? 

[Reeza · Accepted solution]

PDF() works on a row basis, uses only data on that row, not on a column basis so it may not be doing what you want anyways.

Do you have it working for one group? How to generate your poisson probability that's been tested?

If the pdf() function does work, then you can use it in the data step without the group by.

Providing some example data may help.