Avoid modifying input data frames in join when creating indicator (#1434

)
JuliaData · Jun 20, 2018 · 0a3f832 · 0a3f832
1 parent 14f30ad
commit 0a3f832
Show file tree

Hide file tree

Showing 2 changed files with 8 additions and 2 deletions.
diff --git a/src/abstractdataframe/join.jl b/src/abstractdataframe/join.jl
@@ -296,8 +296,8 @@ function Base.join(df1::AbstractDataFrame,
                  indicator_cols[i] *= 'X'
             end
          end
-        df1[Symbol(indicator_cols[1])] = trues(nrow(df1))
-        df2[Symbol(indicator_cols[2])] = trues(nrow(df2))
+         df1 = hcat(df1, DataFrame(Dict(indicator_cols[1] => trues(nrow(df1)))))
+         df2 = hcat(df2, DataFrame(Dict(indicator_cols[2] => trues(nrow(df2)))))
     end
 
     if kind == :cross

diff --git a/test/join.jl b/test/join.jl
@@ -584,8 +584,14 @@ module TestJoin
                                     Name = ["John Doe", "Jane Doe", "Jane Doe", "Joe Blogs", missing],
                                     Job = ["Lawyer", "Doctor", "Florist", missing, "Farmer"],
                                     _merge = ["both", "both", "both", "left_only", "right_only"])
+
+        # Check that input data frame isn't modified (#1434)
+        pre_join_name = copy(name)
+        pre_join_job = copy(job)
         @test join(name, job, on = :ID, kind = :outer, indicator=:_merge,
                    makeunique=true) ≅ outer_indicator
+        @test name ≅ pre_join_name
+        @test job ≅ pre_join_job
 
         # Works with conflicting names
         name2 = DataFrame(ID = [1, 2, 3], Name = ["John Doe", "Jane Doe", "Joe Blogs"],