diff --git a/docs/src/lib/indexing.md b/docs/src/lib/indexing.md
index ed2ca488d..36bd1fd2b 100644
--- a/docs/src/lib/indexing.md
+++ b/docs/src/lib/indexing.md
@@ -33,7 +33,9 @@ The following values are a valid column index:
     * a `Not` expression (see
       [InvertedIndices.jl](https://github.com/JuliaData/InvertedIndices.jl));
       `Not(idx)` selects all indices not in the passed `idx`;
-      when passed as column selector `Not(idx...)` is equivalent to `Not(Cols(idx...))`.
+      when passed as column selector `Not(idx...)` is equivalent to
+      `Not([idx...])` if all indices are integers or to
+      `Not(Cols(idx...))` otherwise.
     * a `Cols` expression (see
       [DataAPI.jl](https://github.com/JuliaData/DataAPI.jl)); `Cols(idxs...)`
       selects the union of the selections in `idxs`; in particular `Cols()`
diff --git a/test/indexing.jl b/test/indexing.jl
index 7660a9601..c7dd69d11 100644
--- a/test/indexing.jl
+++ b/test/indexing.jl
@@ -20,8 +20,11 @@ using Test, DataFrames, Unicode, Random
     end
 
     @test df[!, Not(1, 2)] == DataFrame(c=7:9)
+    @test_throws ArgumentError df[!, Not(1, 1, 2)]
     @test df[!, Not(:b, 1)] == DataFrame(c=7:9)
+    @test df[!, Not(:b, :b, 1)] == DataFrame(c=7:9)
     @test df[!, Not("c", :a)] == DataFrame(b=4:6)
+    @test df[!, Not("c", "c", :a)] == DataFrame(b=4:6)
     @test df[!, Not(:b, "c", :a)] == DataFrame()
     @test df[!, Not([1, 2], :b)] == DataFrame(c=7:9)
     @test df[!, Not([:c, :a], :b)] == DataFrame()