avoid not specialized Pair issue (#2889)

Author: bkamins

NEWS.md

@@ -77,6 +77,8 @@
 
 * fix a problem with `unstack` on empty data frame
   ([#2842](https://github.com/JuliaData/DataFrames.jl/issues/2842))
+* fix a problem with not specialized `Pair` arguments passed as transformations
+  ([#2889](https://github.com/JuliaData/DataFrames.jl/issues/2889))
 
 # DataFrames.jl v1.2.2 Patch Release Notes
 

src/abstractdataframe/selection.jl

@@ -182,6 +182,10 @@ end
 # add a method to funname defined in other/utils.jl
 funname(row::ByRow) = funname(row.fun)
 
+make_pair_concrete(@nospecialize(x::Pair)) =
+    make_pair_concrete(x.first) => make_pair_concrete(x.second)
+make_pair_concrete(@nospecialize(x)) = x
+
 normalize_selection(idx::AbstractIndex, @nospecialize(sel), renamecols::Bool) =
     try
         idx[sel]
@@ -1328,7 +1332,7 @@ function manipulate(df::DataFrame, @nospecialize(cs...); copycols::Bool, keeprow
             push!(cs_vec, v)
         end
     end
-    return _manipulate(df, Any[normalize_selection(index(df), c, renamecols) for c in cs_vec],
+    return _manipulate(df, Any[normalize_selection(index(df), make_pair_concrete(c), renamecols) for c in cs_vec],
                     copycols, keeprows)
 end
 
@@ -1418,7 +1422,8 @@ function manipulate(dfv::SubDataFrame, @nospecialize(args...); copycols::Bool, k
                 push!(cs_vec, v)
             end
         end
-        return _manipulate(dfv, Any[normalize_selection(index(dfv), c, renamecols) for c in cs_vec],
+        return _manipulate(dfv, Any[normalize_selection(index(dfv),
+                                    make_pair_concrete(c), renamecols) for c in cs_vec],
                            true, keeprows)
     else
         # we do not support transformations here
@@ -1436,7 +1441,7 @@ function manipulate(dfv::SubDataFrame, @nospecialize(args...); copycols::Bool, k
                     push!(seen_single_column, ind_idx)
                 end
             else
-                newind = normalize_selection(index(dfv), ind, renamecols)
+                newind = normalize_selection(index(dfv), make_pair_concrete(ind), renamecols)
                 if newind isa Pair
                     throw(ArgumentError("transforming and renaming columns of a " *
                                         "SubDataFrame is not allowed when `copycols=false`"))

src/groupeddataframe/splitapplycombine.jl

@@ -62,7 +62,7 @@ function _combine_prepare_norm(gd::GroupedDataFrame,
     cs_norm = []
     optional_transform = Bool[]
     for c in cs_vec
-        arg = normalize_selection(index(parent(gd)), c, renamecols)
+        arg = normalize_selection(index(parent(gd)), make_pair_concrete(c), renamecols)
         if arg isa AbstractVector{Int}
             for col_idx in arg
                 push!(cs_norm, col_idx => identity => _names(gd)[col_idx])

test/select.jl

@@ -1678,4 +1678,18 @@ end
     @test select(df, :a => ByRow(f) => :a) == DataFrame(a=1:3)
 end
 
+@testset "vectors of pairs with non-specific type are accepted" begin
+    df = DataFrame(x=[1,2,3])
+    @test combine(df, [1 => length => :a, 1 => length => "b"]) == DataFrame(a=3, b=3)
+    @test combine(df, [:x => length => :a, 1 => :b]) == DataFrame(a=3, b=1:3)
+    gdf = groupby(df, :x)
+    @test combine(gdf, [1 => length => :a, 1 => length => "b"]) == DataFrame(x=1:3, a=1, b=1)
+    @test combine(gdf, [:x => length => :a, 1 => :b]) == DataFrame(x=1:3, a=1, b=1:3)
+    sdf = view(df, :, :)
+    @test select(sdf, [1 => length => :a, 1 => length => "b"]) == DataFrame(a=[3, 3, 3], b=[3, 3, 3])
+    @test select(sdf, [:x => length => :a, 1 => :b]) == DataFrame(a=3, b=1:3)
+    @test_throws ArgumentError select(sdf, [1 => length => :a, 1 => length => "b"], copycols=false)
+    @test_throws ArgumentError select(sdf, [:x => length => :a, 1 => :b], copycols=false)
+end
+
 end # module