Fix InexactError in groupby() due to too many levels

Using the storage size of the first PDA does not make sense, since it can have
very few levels, and yet its combination with other columns may produce a lot
of them. UInt32 will be enough for any reasonable number of levels.

src/groupeddataframe/grouping.jl

@@ -90,7 +90,8 @@ function groupby{T}(d::AbstractDataFrame, cols::Vector{T})
     dv = PooledDataArray(d[cols[ncols]])
     # if there are NAs, add 1 to the refs to avoid underflows in x later
     dv_has_nas = (findfirst(dv.refs, 0) > 0 ? 1 : 0)
-    x = copy(dv.refs) .+ dv_has_nas
+    # use UInt32 instead of the PDA's integer size since the number of levels can be high
+    x = copy!(similar(dv.refs, UInt32), dv.refs) .+ dv_has_nas
     # also compute the number of groups, which is the product of the set lengths
     ngroups = length(dv.pool) + dv_has_nas
     # if there's more than 1 column, do roughly the same thing repeatedly

test/grouping.jl

@@ -29,4 +29,9 @@ module TestGrouping
     h(df) = g(f(df))
 
     @test combine(map(h, gd)) == combine(map(g, ga))
+
+    # issue #960
+    x = pool(collect(1:20))
+    df = DataFrame(v1=x, v2=x)
+    groupby(df, [:v1, :v2])
 end