# Applying different functions to Dataset columns after GroupBy

I have a list of associations dataset.

  d = Dataset@{
<|"a" -> "Red", "b" -> .1|>,
<|"a" -> "Blue", "b" -> .2|>,
<|"a" -> "Red", "b" -> .11|>,
<|"a" -> "Blue", "b" -> .21|>,
<|"a" -> "Blue", "b" -> .23|>
}


I can count each type of "a" with

d[GroupBy["a"], Length, "b"]

Red  2
Blue  3


and I can find the mean value of "b" for each "a" with

d[GroupBy["a"], Mean, "b"]

Red  0.105
Blue  0.21333


Is there any easy way I can do both at the same time, getting a result that has both a count of the number of occurrences and the mean "b" for each "a"? I'd want the result to look like

Red 2 0.105
Blue 3 0.21333


Obviously I know how to do the queries separately, and I could figure out how to glue the results together. I also know how to apply different functions to different columns without the GroupBy as in here. And I know how to Map multiple functions to individual entries of a column, creating new columns after Group By. But the GroupBy is messing me up when I try to do Mean or Length on an overall column. Is there some arcane syntax to do this elegantly?

I think I figured it out

d[GroupBy["a"], {Query[Length, "b"], Query[Mean, "b"]}]


although I can't quite explain why that works.

UPDATE -- and it doesn't really work. If you look at the Normal

<|"Red" -> {2, 0.105}, "Blue" -> {3, 0.213333}|>


d[GroupBy["a"], <| "Length" -> Query[Length, "b"],
"Mean" -> Query[Mean, "b"]|>]


But an even better answer is simply

 d[GroupBy["a"], <|"Length" -> Length, "Mean" -> Mean|>, "b"]


Dunno why I didn't see it earlier.

• Very very nice (+1) – eldo Jul 6 '17 at 16:05
• But I don't think this is really the answer. If you look at the Normal of the result, the second column is made up of lists of the two queries. That's not right. The result should have three columns. – Chris Nadovich Jul 6 '17 at 16:17
• The consequences of ordering things in the square brackets are very subtle. – Chris Nadovich Jul 6 '17 at 16:36