Is it possible to compute stats grouped by a column?

[xiaodaigh]

The current GroupBy accepts a type. But is it possible to say, I have data coming in that is of the form (grp, x, y) And I wish to compute the same stats for x and y but I create a set of stats for each distinct value of grp?

Like a group by statement in SQL?

select
  mean(x),
  mean(y),
from
  data
group by
  grp;

[joshday]

If I'm understanding correctly, you would use a Series as the stat in your GroupBy:

GroupBy(typeof(grp), Series(Mean(), Mean()))

Edit:

Note that this would expect an iterator with values like (grp, (x,y))

[xiaodaigh]

Ok. what if grp and x have the same type? Or something like that. Or I need to wrap grp in its own type.