diff --git a/src/groupeddataframe/grouping.jl b/src/groupeddataframe/grouping.jl
index 085a685059..6b1d7f050b 100644
--- a/src/groupeddataframe/grouping.jl
+++ b/src/groupeddataframe/grouping.jl
@@ -170,12 +170,13 @@ wrap(s::Any) = DataFrame(x1 = s)
 Combine a GroupApplied object (rudimentary)
 
 ```julia
-combine(ga::GroupApplied)
+combine(ga::GroupApplied; append_keys::Bool=false)
 ```
 
 ### Arguments
 
 * `ga` : a GroupApplied
+* `append_keys` : whether or not to append the grouped by columns.
 
 ### Returns
 
@@ -192,7 +193,7 @@ combine(map(d -> mean(skipmissing(d[:c])), gd))
 ```
 
 """
-function combine(ga::GroupApplied)
+function combine(ga::GroupApplied; append_keys::Bool=false)
     gd, vals = ga.gd, ga.vals
     valscat = _vcat(vals)
     idx = Vector{Int}(undef, size(valscat, 1))
@@ -202,7 +203,7 @@ function combine(ga::GroupApplied)
         idx[j .+ (1:n)] .= gd.idx[start]
         j += n
     end
-    hcat!(gd.parent[idx, gd.cols], valscat)
+    append_keys ? hcat!(gd.parent[idx, gd.cols], valscat) : valscat
 end
 
 
@@ -250,8 +251,8 @@ Split-apply-combine in one step; apply `f` to each grouping in `d`
 based on columns `col`
 
 ```julia
-by(d::AbstractDataFrame, cols, f::Function; sort::Bool = false)
-by(f::Function, d::AbstractDataFrame, cols; sort::Bool = false)
+by(d::AbstractDataFrame, cols, f::Function; sort::Bool=false, append_keys::Bool=false)
+by(f::Function, d::AbstractDataFrame, cols; sort::Bool=false, append_keys::Bool=false)
 ```
 
 ### Arguments
@@ -261,6 +262,7 @@ by(f::Function, d::AbstractDataFrame, cols; sort::Bool = false)
 * `f` : a function to be applied to groups; expects each argument to
   be an AbstractDataFrame
 * `sort`: sort row groups (no sorting by default)
+* `append_keys`: whether or not to include the columns that the DataFrame was grouped by
 
 `f` can return a value, a vector, or a DataFrame. For a value or
 vector, these are merged into a column along with the `cols` keys. For
@@ -293,10 +295,10 @@ end
 ```
 
 """
-by(d::AbstractDataFrame, cols, f::Function; sort::Bool = false) =
-    combine(map(f, groupby(d, cols, sort = sort)))
-by(f::Function, d::AbstractDataFrame, cols; sort::Bool = false) =
-    by(d, cols, f, sort = sort)
+by(d::AbstractDataFrame, cols, f::Function; sort::Bool=false, append_keys::Bool=false) =
+    combine(map(f, groupby(d, cols, sort=sort)), append_keys=append_keys)
+by(f::Function, d::AbstractDataFrame, cols; sort::Bool=false, append_keys::Bool=false) =
+    by(d, cols, f, sort=sort, append_keys=append_keys)
 
 #
 # Aggregate convenience functions
@@ -349,7 +351,7 @@ end
 aggregate(gd::GroupedDataFrame, f::Function; sort::Bool=false) = aggregate(gd, [f], sort=sort)
 function aggregate(gd::GroupedDataFrame, fs::Vector{T}; sort::Bool=false) where T<:Function
     headers = _makeheaders(fs, setdiff(_names(gd), _names(gd.parent[gd.cols])))
-    res = combine(map(x -> _aggregate(without(x, gd.cols), fs, headers), gd))
+    res = combine(map(x -> _aggregate(without(x, gd.cols), fs, headers), gd), append_keys=true)
     sort && sort!(res, headers)
     res
 end
diff --git a/test/grouping.jl b/test/grouping.jl
index be2643ad82..900c9b68c8 100644
--- a/test/grouping.jl
+++ b/test/grouping.jl
@@ -92,30 +92,30 @@ module TestGrouping
     sdf = unique(df[cols])
 
     # by() without groups sorting
-    bdf = by(df, cols, f)
+    bdf = by(df, cols, f, append_keys=true)
     @test bdf[cols] == sdf
 
     # by() with groups sorting
-    sbdf = by(df, cols, f, sort=true)
+    sbdf = by(df, cols, f, sort=true, append_keys=true)
     @test sbdf[cols] == sort(sdf)
 
-    byf = by(df, :a, df -> DataFrame(bsum = sum(df[:b])))
+    byf = by(df, :a, df -> DataFrame(bsum = sum(df[:b])), append_keys=true)
 
     # groupby() without groups sorting
     gd = groupby(df, cols)
     ga = map(f, gd)
 
-    @test bdf == combine(ga)
+    @test bdf == combine(ga, append_keys=true)
 
     # groupby() with groups sorting
     gd = groupby(df, cols, sort=true)
     ga = map(f, gd)
-    @test sbdf == combine(ga)
+    @test sbdf == combine(ga, append_keys=true)
 
     g(df) = DataFrame(cmax1 = [c + 1 for c in df[:cmax]])
     h(df) = g(f(df))
 
-    @test combine(map(h, gd)) == combine(map(g, ga))
+    @test combine(map(h, gd), append_keys=true) == combine(map(g, ga), append_keys=true)
 
     # testing pool overflow
     df2 = DataFrame(v1 = categorical(collect(1:1000)), v2 = categorical(fill(1, 1000)))
@@ -142,7 +142,7 @@ module TestGrouping
     df = DataFrame(v1=x, v2=x)
     groupby(df, [:v1, :v2])
 
-    df2 = by(e->1, DataFrame(x=Int64[]), :x)
+    df2 = by(e->1, DataFrame(x=Int64[]), :x, append_keys=true)
     @test size(df2) == (0, 1)
     @test sum(df2[:x]) == 0