From 1a241c953c3357830df79d0372178827d60bc25b Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Bogumi=C5=82=20Kami=C5=84ski?= <bkamins@sgh.waw.pl>
Date: Fri, 13 Apr 2018 19:43:02 +0200
Subject: [PATCH] fix issorted bug (#1391)

---
 src/abstractdataframe/sort.jl | 17 +++++++-----
 test/sort.jl                  | 50 +++++++++++++++++++++++++++++++++++
 2 files changed, 60 insertions(+), 7 deletions(-)

diff --git a/src/abstractdataframe/sort.jl b/src/abstractdataframe/sort.jl
index 22bd23e4c6..20197ec2f2 100644
--- a/src/abstractdataframe/sort.jl
+++ b/src/abstractdataframe/sort.jl
@@ -73,14 +73,11 @@ DFPerm(o::O, df::DF) where {O<:Ordering, DF<:AbstractDataFrame} = DFPerm{O,DF}(o
 col_ordering(o::DFPerm{O}, i::Int) where {O<:Ordering} = o.ord
 col_ordering(o::DFPerm{V}, i::Int) where {V<:AbstractVector} = o.ord[i]
 
-Base.@propagate_inbounds Base.getindex(o::DFPerm, i::Int, j::Int) = o.df[i, j]
-Base.@propagate_inbounds Base.getindex(o::DFPerm, a::DataFrameRow, j::Int) = a[j]
-
 function Sort.lt(o::DFPerm, a, b)
-    @inbounds for i = 1:ncol(o.df)
+    @inbounds for i in 1:ncol(o.df)
         ord = col_ordering(o, i)
-        va = o[a, i]
-        vb = o[b, i]
+        va = o.df[a, i]
+        vb = o.df[b, i]
         lt(ord, va, vb) && return true
         lt(ord, vb, va) && return false
     end
@@ -272,7 +269,13 @@ function Base.issorted(df::AbstractDataFrame, cols_new=[]; cols=[],
                      :issorted)
         cols_new = cols
     end
-    issorted(eachrow(df), ordering(df, cols_new, lt, by, rev, order))
+    if cols_new isa ColumnIndex
+        issorted(df[cols_new], lt=lt, by=by, rev=rev, order=order)
+    elseif length(cols_new) == 1
+        issorted(df[cols_new[1]], lt=lt, by=by, rev=rev, order=order)
+    else
+        issorted(1:nrow(df), ordering(df, cols_new, lt, by, rev, order))
+    end
 end
 
 # sort and sortperm functions
diff --git a/test/sort.jl b/test/sort.jl
index e1bb06cc57..2aa1d50e28 100644
--- a/test/sort.jl
+++ b/test/sort.jl
@@ -42,6 +42,17 @@ module TestSort
 
     @test df == ds
 
+    df = DataFrame(x = [3, 1, 2, 1], y = ["b", "c", "a", "b"])
+    @test !issorted(df, :x)
+    @test issorted(sort(df, :x), :x)
+
+    x = DataFrame(a=1:3,b=3:-1:1,c=3:-1:1)
+    @test issorted(x)
+    @test !issorted(x, [:b,:c])
+    @test !issorted(x[2:3], [:b,:c])
+    @test issorted(sort(x,[2,3]), [:b,:c])
+    @test issorted(sort(x[2:3]), [:b,:c])
+
     # Check that columns that shares the same underlying array are only permuted once PR#1072
     df = DataFrame(a=[2,1])
     df[:b] = df[:a]
@@ -52,4 +63,43 @@ module TestSort
     sort!(x, :y)
     @test x[:y] == [1,2,3,4]
     @test x[:x] == [1,3,2,4]
+
+    @test_throws ArgumentError sort(x, by=:x)
+
+    srand(1)
+    # here there will be probably no ties
+    df_rand1 = DataFrame(rand(100, 4))
+    # but here we know we will have ties
+    df_rand2 = df_rand1[:]
+    df_rand2[:x1] = shuffle([fill(1, 50); fill(2, 50)])
+    df_rand2[:x4] = shuffle([fill(1, 50); fill(2, 50)])
+
+    # test sorting by 1 column
+    for df_rand in [df_rand1, df_rand2]
+        # testing sort
+        for n1 in names(df_rand)
+            # passing column name
+            @test sort(df_rand, n1) == df_rand[sortperm(df_rand[n1]),:]
+            # passing vector with one column name
+            @test sort(df_rand, [n1]) == df_rand[sortperm(df_rand[n1]),:]
+            # passing vector with two column names
+            for n2 in setdiff(names(df_rand), [n1])
+                @test sort(df_rand, [n1,n2]) == df_rand[sortperm(collect(zip(df_rand[n1],
+                                                                             df_rand[n2]))),:]
+            end
+        end
+        # testing if sort! is consistent with issorted and sort
+        ref_df = df_rand[:]
+        for n1 in names(df_rand)
+            @test sort!(df_rand, n1) == sort(ref_df, n1)
+            @test issorted(df_rand, n1)
+            @test sort!(df_rand, [n1]) == sort(ref_df, [n1])
+            @test issorted(df_rand, [n1])
+            for n2 in setdiff(names(df_rand), [n1])
+                @test sort!(df_rand, [n1, n2]) == sort(ref_df, [n1, n2])
+                @test issorted(df_rand, n1)
+                @test issorted(df_rand, [n1, n2])
+            end
+        end
+    end
 end