daru 0.0.5 → 0.1.0

data/lib/daru/version.rb

@@ -0,0 +1,3 @@
+module Daru
+  VERSION = "0.1.0"
+end

data/spec/accessors/wrappers_spec.rb

@@ -33,4 +33,55 @@ end
 
 describe Daru::Accessors::ArrayWrapper do
   
+end
+
+describe Daru::Accessors::GSLWrapper do
+  before :each do
+    @stub_context = Object.new
+    @gsl_wrapper = Daru::Accessors::GSLWrapper.new([1,2,3,4,5,6], @stub_context)
+  end
+
+  context ".new" do
+    it "actually creates a GSL Vector" do
+      expect(@gsl_wrapper.data.class).to eq(GSL::Vector)
+    end
+  end
+
+  context "#mean" do
+    it "computes mean" do
+      expect(@gsl_wrapper.mean).to eq(3.5)
+    end
+  end
+
+  context "#map!" do
+    it "destructively maps" do
+      expect(@gsl_wrapper.map! { |a| a += 1 }).to eq(
+        Daru::Accessors::GSLWrapper.new([2,3,4,5,6,7], @stub_context)
+      )
+    end
+  end
+
+  context "#delete_at" do
+    it "deletes at key" do
+      expect(@gsl_wrapper.delete_at(2)).to eq(3)
+
+      expect(@gsl_wrapper).to eq(
+        Daru::Accessors::GSLWrapper.new([1,2,4,5,6], @stub_context)
+      )
+    end
+  end
+
+  context "#index" do
+    it "returns index of value" do
+      expect(@gsl_wrapper.index(3)).to eq(2)
+    end
+  end
+
+  context "#push" do
+    it "appends element" do
+      expect(@gsl_wrapper.push(15)).to eq(
+        Daru::Accessors::GSLWrapper.new([1,2,3,4,5,6,15], @stub_context)
+      )
+    end
+  end
 end

data/spec/core/group_by_spec.rb

@@ -96,8 +96,6 @@ describe Daru::Core::GroupBy do
         d: [22]
         }, index: [1]
       ))
-
-      puts @dl_group.get_group(["foo", "one"]).inspect
     end
 
     it "returns the whole sub-group for triple layer grouping" do

data/spec/daru_spec.rb

@@ -0,0 +1,58 @@
+require 'spec_helper.rb'
+
+describe "Daru.lazy_update" do
+  context "A variable which will set whether Vector metadata is updated immediately or lazily." do
+    describe Daru::Vector do
+      it "does updates metadata immediately when LAZY_UPDATE is set to default false" do
+        v = Daru::Vector.new [1,2,3,4,nil,nil,3,nil]
+        v[1] = nil
+
+        expect(v.missing_positions.include?(1)).to eq(true)
+      end
+
+      it "does NOT update metadata immediately when @@lazy_update is set to default true. Update done when #update is called" do
+        Daru.lazy_update = true
+        v    = Daru::Vector.new [1,2,3,4,nil,nil]
+        v[1] = nil
+        v[0] = nil
+
+        expect(v.missing_positions.include?(0)).to eq(false)
+        expect(v.missing_positions.include?(1)).to eq(false)
+
+        v.update
+        expect(v.missing_positions.include?(0)).to eq(true)
+        expect(v.missing_positions.include?(1)).to eq(true)
+
+        Daru.lazy_update = false
+      end
+    end
+
+    describe Daru::DataFrame do
+      before do 
+        v = Daru::Vector.new [1,2,3,4,nil,nil,3,nil]
+        @df = Daru::DataFrame.new({a: v, b: v, c: v})
+      end
+
+      it "does updates metadata immediately when LAZY_UPDATE is set to default false" do
+        @df[:a][1] = nil
+
+        expect(@df[:a].missing_positions.include?(1)).to eq(true)
+      end
+
+      it "does NOT update metadata immediately when @@lazy_update is set to default true. Update done when #update is called" do
+        Daru.lazy_update = true
+        @df[:c][0] = nil
+        @df[:a][1] = nil
+
+        expect(@df[:c].missing_positions.include?(0)).to eq(false)
+        expect(@df[:a].missing_positions.include?(1)).to eq(false)
+
+        @df.update
+        expect(@df[:c].missing_positions.include?(0)).to eq(true)
+        expect(@df[:a].missing_positions.include?(1)).to eq(true)
+
+        Daru.lazy_update = false
+      end      
+    end
+  end
+end

data/spec/dataframe_spec.rb

@@ -264,6 +264,37 @@ describe Daru::DataFrame do
         expect(df.vectors).to eq([:a,:c,:b].to_index)
       end
 
+      it "does not copy vectors when clone: false" do
+        a = Daru::Vector.new([1,2,3,4,5])
+        b = Daru::Vector.new([1,2,3,4,5])
+        c = Daru::Vector.new([1,2,3,4,5])
+        df = Daru::DataFrame.new({a: a, b: b, c: c}, clone: false)
+        
+        expect(df[:a].object_id).to eq(a.object_id)
+        expect(df[:b].object_id).to eq(b.object_id)
+        expect(df[:c].object_id).to eq(c.object_id)
+      end
+
+      it "allows creation of empty dataframe with only order", focus: true do
+        df = Daru::DataFrame.new({}, order: [:a, :b, :c])
+        df[:a] = Daru::Vector.new([1,2,3,4,5,6])
+
+        expect(df.size).to eq(6)
+        expect(df[:a]).to eq(Daru::Vector.new([1,2,3,4,5,6]))
+        expect(df[:b]).to eq(Daru::Vector.new([nil,nil,nil,nil,nil,nil]))
+        expect(df[:c]).to eq(Daru::Vector.new([nil,nil,nil,nil,nil,nil]))
+      end
+
+      it "allows creation of dataframe without specifying order or index" do
+        df = Daru::DataFrame.new({})
+        df[:a] = Daru::Vector.new([1,2,3,4,5])
+
+        expect(df.size)        .to eq(5)
+        expect(df.index.to_a)  .to eq([0,1,2,3,4])
+        expect(df.vectors.to_a).to eq([:a])
+        expect(df[:a])         .to eq(Daru::Vector.new([1,2,3,4,5]))
+      end
+
       it "raises error for incomplete DataFrame index" do
         expect {
           df = Daru::DataFrame.new({b: [11,12,13,14,15], a: [1,2,3,4,5], 
@@ -390,6 +421,14 @@ describe Daru::DataFrame do
       it "accesses vector with Integer index" do
         expect(@df[0, :vector]).to eq([1,2,3,4,5].dv(:a, [:one, :two, :three, :four, :five]))
       end
+
+      it "returns a subset of DataFrame when specified range" do
+        subset = @df[:b..:c]
+        expect(subset).to eq(Daru::DataFrame.new({
+          b: [11,12,13,14,15],
+          c: [11,22,33,44,55]
+          }, index: [:one, :two, :three, :four, :five]))
+      end
     end
 
     context Daru::MultiIndex do
@@ -501,7 +540,42 @@ describe Daru::DataFrame do
     end
 
     context Daru::MultiIndex do
-      pending
+      it "raises error when incomplete index specified but index is absent" do
+        expect {
+          @df_mi[:d] = [100,200,300,400,100,200,300,400,100,200,300,400]
+        }.to raise_error
+      end
+
+      it "assigns all sub-indexes when a top level index is specified" do
+        pending
+        @df_mi[:a] = [100,200,300,400,100,200,300,400,100,200,300,400]
+        
+        expect(@df_mi).to eq(Daru::DataFrame.new([
+          [100,200,300,400,100,200,300,400,100,200,300,400],
+          [100,200,300,400,100,200,300,400,100,200,300,400],
+          @vector_arry1,
+          @vector_arry2], index: @multi_index, order: @order_mi))  
+      end
+
+      it "creates a new vector when full index specfied" do
+        pending
+        order = Daru::MultiIndex.new([
+          [:a,:one,:bar],
+          [:a,:two,:baz],
+          [:b,:two,:foo],
+          [:b,:one,:foo],
+          [:c,:one,:bar]])
+        answer = Daru::DataFrame.new([
+          @vector_arry1, 
+          @vector_arry2,
+          @vector_arry1,
+          @vector_arry2,
+          [100,200,300,400,100,200,300,400,100,200,300,400]
+          ], index: @multi_index, order: order)
+        @df_mi[:c,:one,:bar] = [100,200,300,400,100,200,300,400,100,200,300,400]
+
+        expect(@df_mi).to eq(answer)
+      end
     end
   end
 
@@ -551,6 +625,12 @@ describe Daru::DataFrame do
         expect(@df.row[:two]).to eq([2,9,11].dv(:two, [:a, :b, :c]))
       end
 
+      it "correctlu aligns assinged DV by index for new rows" do
+        @df.row[:latest] = Daru::Vector.new([2,3,1], index: [:b,:c,:a])
+
+        expect(@df.row[:latest]).to eq(Daru::Vector.new([1,2,3], index: [:a,:b,:c]))
+      end
+
       it "inserts nils for indexes that dont exist in the DataFrame" do
         @df.row[:two] = [49, 99, 59].dv(nil, [:oo, :aah, :gaah])
 
@@ -667,6 +747,19 @@ describe Daru::DataFrame do
     end
   end
 
+  context "#add_row" do
+    it "allows adding rows after making empty DF by specfying only order" do
+      df = Daru::DataFrame.new({}, order: [:a, :b, :c])
+      df.add_row [1,2,3]
+      df.add_row [5,6,7]
+
+      expect(df[:a]).to eq(Daru::Vector.new([1,5]))
+      expect(df[:b]).to eq(Daru::Vector.new([2,6]))
+      expect(df[:c]).to eq(Daru::Vector.new([3,7]))
+      expect(df.index).to eq(Daru::Index.new([0,1]))
+    end
+  end
+
   context "#vector" do
     context Daru::Index do
       it "appends an Array as a Daru::Vector" do
@@ -722,38 +815,89 @@ describe Daru::DataFrame do
   end
 
   context "#dup" do
-    it "dups every data structure inside DataFrame" do
-      clo = @data_frame.dup
+    context Daru::Index do
+      it "dups every data structure inside DataFrame" do
+        clo = @data_frame.dup
 
-      expect(clo.object_id)        .not_to eq(@data_frame.object_id)
-      expect(clo.vectors.object_id).not_to eq(@data_frame.object_id)
-      expect(clo.index.object_id)  .not_to eq(@data_frame.object_id)
+        expect(clo.object_id)        .not_to eq(@data_frame.object_id)
+        expect(clo.vectors.object_id).not_to eq(@data_frame.object_id)
+        expect(clo.index.object_id)  .not_to eq(@data_frame.object_id)
 
-      @data_frame.each_vector_with_index do |vector, index|
-        expect(vector.object_id).not_to eq(clo.vector[index].object_id)
+        @data_frame.each_vector_with_index do |vector, index|
+          expect(vector.object_id).not_to eq(clo.vector[index].object_id)
+        end
       end
     end
-  end
 
-  context "#each_vector" do
-    context Daru::Index do
-      it "iterates over all vectors" do
-        ret = @data_frame.each_vector do |vector|
-          expect(vector.index).to eq([:one, :two, :three, :four, :five].to_index)
-          expect(vector.class).to eq(Daru::Vector) 
-        end
+    context Daru::MultiIndex do
+      it "duplicates with multi index" do
+        clo = @df_mi.dup
 
-        expect(ret).to eq(@data_frame)
+        expect(clo)                  .to     eq(@df_mi)
+        expect(clo.vectors.object_id).not_to eq(@df_mi.vectors.object_id)
+        expect(clo.index.object_id)  .not_to eq(@df_mi.index.object_id)
       end
+    end
+  end
 
-      it "returns Enumerable if no block specified" do
-        ret = @data_frame.each_vector
-        expect(ret.is_a?(Enumerator)).to eq(true) 
-      end
+  context "#dup_only_valid" do
+    before do
+      @missing_data_df = Daru::DataFrame.new({
+        a: [1  , 2, 3, nil, 4, nil, 5],
+        b: [nil, 2, 3, nil, 4, nil, 5],
+        c: [1,   2, 3, 43 , 4, nil, 5]
+      })      
     end
 
-    context Daru::MultiIndex do
+    it "dups rows with non-missing data only" do
+      df = Daru::DataFrame.new({
+        a: [2, 3, 4, 5],
+        b: [2, 3, 4, 5],
+        c: [2, 3, 4, 5]
+      }, index: [1,2,4,6]) 
+      expect(@missing_data_df.dup_only_valid).to eq(df)
+    end
+
+    it "dups only the specified vectors" do
+      df = Daru::DataFrame.new({
+        a: [2,3,4,5],
+        c: [2,3,4,5]
+      }, index: [1,2,4,6])
+      expect(@missing_data_df.dup_only_valid([:a, :c])).to eq(df)
+    end
+  end
+
+  context "#clone" do
+    it "returns a view of the whole dataframe" do
+      cloned = @data_frame.clone
+      expect(@data_frame.object_id).to_not eq(cloned.object_id)
+      expect(@data_frame[:a].object_id).to eq(cloned[:a].object_id)
+      expect(@data_frame[:b].object_id).to eq(cloned[:b].object_id)
+      expect(@data_frame[:c].object_id).to eq(cloned[:c].object_id)
+    end
+
+    it "returns a view of selected vectors" do
+      cloned = @data_frame.clone(:a, :b)
+      expect(cloned.object_id).to_not eq(@data_frame.object_id)
+      expect(cloned[:a].object_id).to eq(@data_frame[:a].object_id)
+      expect(cloned[:b].object_id).to eq(@data_frame[:b].object_id)
+    end
 
+    it "clones properly when supplied array" do
+      cloned = @data_frame.clone([:a, :b])
+      expect(cloned.object_id).to_not eq(@data_frame.object_id)
+      expect(cloned[:a].object_id).to eq(@data_frame[:a].object_id)
+      expect(cloned[:b].object_id).to eq(@data_frame[:b].object_id)
+    end
+  end
+
+  context "#clone_structure" do
+    it "clones only the index and vector structures of the data frame" do
+      cs = @data_frame.clone_structure
+
+      expect(cs.vectors).to eq(@data_frame.vectors)
+      expect(cs.index).to eq(@data_frame.index)
+      expect(cs[:a]).to eq(Daru::Vector.new([nil] * cs[:a].size, index: @data_frame.index))
     end
   end
 
@@ -772,17 +916,6 @@ describe Daru::DataFrame do
     end
   end
 
-  context "#each_row" do
-    it "iterates over rows" do
-      ret = @data_frame.each_row do |row|
-        expect(row.index).to eq([:a, :b, :c].to_index)
-        expect(row.class).to eq(Daru::Vector)
-      end
-
-      expect(ret).to eq(@data_frame)
-    end
-  end
-
   context "#each_row_with_index" do
     it "iterates over rows with indexes" do
       idxs = []
@@ -797,81 +930,154 @@ describe Daru::DataFrame do
     end
   end
 
-  context "#map_vectors" do
-    it "iterates over vectors and returns a modified DataFrame" do
-      ans = Daru::DataFrame.new({b: [21,22,23,24,25], a: [11,12,13,14,15], 
-      c: [21,32,43,54,65]}, order: [:a, :b, :c], 
-      index: [:one, :two, :three, :four, :five])
+  context "#each" do
+    it "iterates over rows" do
+      ret = @data_frame.each(:row) do |row|
+        expect(row.index).to eq([:a, :b, :c].to_index)
+        expect(row.class).to eq(Daru::Vector)
+      end
 
-      ret = @data_frame.map_vectors do |vector|
-        vector = vector.map { |e| e += 10}
+      expect(ret).to eq(@data_frame)
+    end
+
+    it "iterates over all vectors" do
+      ret = @data_frame.each do |vector|
+        expect(vector.index).to eq([:one, :two, :three, :four, :five].to_index)
+        expect(vector.class).to eq(Daru::Vector) 
       end
 
-      expect(ret).to eq(ans)
-      expect(ret == @data_frame).to eq(false)
+      expect(ret).to eq(@data_frame)
+    end
+
+    it "returns Enumerable if no block specified" do
+      ret = @data_frame.each
+      expect(ret.is_a?(Enumerator)).to eq(true) 
     end
   end
 
-  context "#map_vectors!" do
-    it "maps vectors (bang)" do
-      ans = Daru::DataFrame.new({b: [21,22,23,24,25], a: [11,12,13,14,15], 
+  context "#recode" do
+    before do
+      @ans_vector = Daru::DataFrame.new({b: [21,22,23,24,25], a: [11,12,13,14,15], 
         c: [21,32,43,54,65]}, order: [:a, :b, :c], 
         index: [:one, :two, :three, :four, :five])
 
-      @data_frame.map_vectors! do |vector|
+      @ans_rows = Daru::DataFrame.new({b: [121, 144, 169, 196, 225], a: [1,4,9,16,25], 
+        c: [121, 484, 1089, 1936, 3025]}, order: [:a, :b, :c], 
+        index: [:one, :two, :three, :four, :five])
+    end
+
+    it "maps over the vectors of a DataFrame and returns a DataFrame" do
+      ret = @data_frame.recode do |vector|
         vector.map! { |e| e += 10}
       end
 
-      expect(@data_frame).to eq(ans)
+      expect(ret).to eq(@ans_vector)
+    end
+
+    it "maps over the rows of a DataFrame and returns a DataFrame" do
+      ret = @data_frame.recode(:row) do |row|
+        expect(row.class).to eq(Daru::Vector)
+        row.map! { |e| e*e }
+      end
+
+      expect(ret).to eq(@ans_rows)
     end
   end
 
-  context "#map_vectors_with_index" do
-    it "iterates over vectors with index and returns a modified DataFrame" do
-      ans = Daru::DataFrame.new({b: [21,22,23,24,25], a: [11,12,13,14,15], 
-      c: [21,32,43,54,65]}, order: [:a, :b, :c], 
-      index: [:one, :two, :three, :four, :five])
+  context "#collect" do
+    before do
+      @df = Daru::DataFrame.new({
+        a: [1,2,3,4,5], 
+        b: [11,22,33,44,55],
+        c: [1,2,3,4,5]
+      })
+    end
 
-      idx = []
-      ret = @data_frame.map_vectors_with_index do |vector, index|
-        idx << index
-        vector = vector.map { |e| e += 10}
+    it "collects calculation over rows and returns a Vector from the results" do
+      expect(@df.collect(:row) { |row| (row[:a] + row[:c]) * row[:c] }).to eq(
+        Daru::Vector.new([2,8,18,32,50])
+        )
+    end
+
+    it "collects calculation over vectors and returns a Vector from the results" do
+      expect(@df.collect { |v| v[0] * v[1] + v[4] }).to eq(
+        Daru::Vector.new([7,297,7], index: [:a, :b, :c])
+        )
+    end
+  end
+
+  context "#map" do
+    it "iterates over rows and returns an Array" do
+      ret = @data_frame.map(:row) do |row|
+        expect(row.class).to eq(Daru::Vector)
+        row[:a] * row[:c]
       end
 
-      expect(ret).to eq(ans)
-      expect(idx).to eq([:a, :b, :c])
+      expect(ret).to eq([11, 44, 99, 176, 275])
+      expect(@data_frame.vectors.to_a).to eq([:a, :b, :c])
+    end
+
+    it "iterates over vectors and returns an Array" do
+      ret = @data_frame.map do |vector|
+        vector.mean
+      end
+      expect(ret).to eq([3.0, 13.0, 33.0])
     end
   end
 
-  context "#map_rows" do
-    it "iterates over rows and returns a modified DataFrame" do
-      ans = Daru::DataFrame.new({b: [121, 144, 169, 196, 225], a: [1,4,9,16,25], 
-        c: [121, 484, 1089, 1936, 3025]}, order: [:a, :b, :c], 
+  context "#map!" do
+    before do
+      @ans_vector = Daru::DataFrame.new({b: [21,22,23,24,25], a: [11,12,13,14,15], 
+        c: [21,32,43,54,65]}, order: [:a, :b, :c], 
         index: [:one, :two, :three, :four, :five])
 
-      ret = @data_frame.map_rows do |row|
-        expect(row.class).to eq(Daru::Vector)
-        row = row.map { |e| e*e }
+      @ans_row = Daru::DataFrame.new({b: [12,13,14,15,16], a: [2,3,4,5,6], 
+        c: [12,23,34,45,56]}, order: [:a, :b, :c], 
+        index: [:one, :two, :three, :four, :five])
+    end
+
+    it "destructively maps over the vectors and changes the DF" do
+      @data_frame.map! do |vector|
+        vector + 10
+      end
+      expect(@data_frame).to eq(@ans_vector)
+    end
+
+    it "destructively maps over the rows and changes the DF" do
+      @data_frame.map!(:row) do |row|
+        row + 1
       end
 
-      expect(ret).to eq(ans)
+      expect(@data_frame).to eq(@ans_row)
+    end
+  end
+
+  context "#map_vectors_with_index" do
+    it "iterates over vectors with index and returns an Array" do
+      idx = []
+      ret = @data_frame.map_vectors_with_index do |vector, index|
+        idx << index
+        vector.recode { |e| e += 10}
+      end
+
+      expect(ret).to eq([
+        Daru::Vector.new([11,12,13,14,15],index: [:one, :two, :three, :four, :five]),
+        Daru::Vector.new([21,22,23,24,25],index: [:one, :two, :three, :four, :five]), 
+        Daru::Vector.new([21,32,43,54,65],index: [:one, :two, :three, :four, :five])])
+      expect(idx).to eq([:a, :b, :c])
     end
   end
 
   context "#map_rows_with_index" do
     it "iterates over rows with index and returns a modified DataFrame" do
-      ans = Daru::DataFrame.new({b: [121, 144, 169, 196, 225], a: [1,4,9,16,25], 
-        c: [121, 484, 1089, 1936, 3025]},order: [:a, :b, :c], 
-        index: [:one, :two, :three, :four, :five])
-
       idx = []
       ret = @data_frame.map_rows_with_index do |row, index|
         idx << index
         expect(row.class).to eq(Daru::Vector)
-        row = row.map { |e| e*e }
+        row[:a] * row[:c]
       end
 
-      expect(ret).to eq(ans)
+      expect(ret).to eq([11, 44, 99, 176, 275])
       expect(idx).to eq([:one, :two, :three, :four, :five])
     end
   end
@@ -926,6 +1132,23 @@ describe Daru::DataFrame do
     end
   end
 
+  context "#filter_field" do
+    before do
+      @df = Daru::DataFrame.new({ 
+        :id => Daru::Vector.new([1, 2, 3, 4, 5]), 
+        :name => Daru::Vector.new(%w(Alex Claude Peter Franz George)), 
+        :age => Daru::Vector.new([20, 23, 25, 27, 5]),
+        :city => Daru::Vector.new(['New York', 'London', 'London', 'Paris', 'Tome']),
+        :a1 => Daru::Vector.new(['a,b', 'b,c', 'a', nil, 'a,b,c']) },
+        order: [:id, :name, :age, :city, :a1])
+    end
+
+    it "creates new vector with the data of a given field for which block returns true" do
+      filtered = @df.filter_vector(:id) { |c| c[:id] == 2 or c[:id] == 4 }
+      expect(filtered).to eq(Daru::Vector.new([2,4]))
+    end
+  end
+
   context "#filter_rows" do
     context Daru::Index do
       it "filters rows" do
@@ -988,6 +1211,21 @@ describe Daru::DataFrame do
     end
   end
 
+  context "#to_hash" do
+    it "converts to a hash" do
+      expect(@data_frame.to_hash).to eq( 
+        {
+          a: Daru::Vector.new([1,2,3,4,5], 
+            index: [:one, :two, :three, :four, :five]), 
+          b: Daru::Vector.new([11,12,13,14,15], 
+            index: [:one, :two, :three, :four, :five]),
+          c: Daru::Vector.new([11,22,33,44,55], 
+            index: [:one, :two, :three, :four, :five])
+        }
+      )
+    end
+  end
+
   context "#recast" do
     it "recasts underlying vectors" do
       @data_frame.recast a: :nmatrix, c: :nmatrix
@@ -1137,6 +1375,28 @@ describe Daru::DataFrame do
     end  
   end
 
+  context "#reindex_vectors!" do
+    before :each do
+      @df = Daru::DataFrame.new({
+        a: [1,2,3,4,5],
+        b: [11,22,33,44,55],
+        c: %w(a b c d e)
+      })
+    end
+
+    it "changes names of vectors" do
+      ans = Daru::DataFrame.new({
+        a: [1,2,3,4,5],
+        b: [11,22,33,44,55],
+        c: %w(a b c d e)
+      }, order: [:b, :c, :a])
+
+      expect(@df.reindex_vectors!([:b,:c,:a])).to eq(Daru::Index.new([:b,:c,:a], [1,2,0]))
+      expect(@df[:a]).to eq(Daru::Vector.new([1,2,3,4,5]))
+      expect(@df[:c]).to eq(Daru::Vector.new(%w(a b c d e)))
+    end
+  end
+
   context "#to_matrix" do
     before do
       @df = Daru::DataFrame.new({b: [11,12,13,14,15], a: [1,2,3,4,5], 
@@ -1292,7 +1552,7 @@ describe Daru::DataFrame do
       ))
     end
 
-    it "creates row and vector index with (double) index and (double) vector args", focus: true do
+    it "creates row and vector index with (double) index and (double) vector args" do
       agg_index = Daru::MultiIndex.new([
         [:bar, 4],
         [:bar, 5],
@@ -1394,4 +1654,439 @@ describe Daru::DataFrame do
       expect(@data_frame.shape).to eq([5,3])
     end
   end
+
+  context "#nest" do
+    it "nests in a hash" do
+      df = Daru::DataFrame.new({
+        :a => Daru::Vector.new(%w(a a a b b b)),
+        :b => Daru::Vector.new(%w(c c d d e e)),
+        :c => Daru::Vector.new(%w(f g h i j k))
+      })
+      nest = df.nest :a, :b
+      expect(nest['a']['c']).to eq([{ :c => 'f' }, { :c => 'g' }])
+      expect(nest['a']['d']).to eq([{ :c => 'h' }])
+      expect(nest['b']['e']).to eq([{ :c => 'j' }, { :c => 'k' }])
+    end
+  end
+
+  context "#summary" do
+    it "produces a summary of data frame" do
+      expect(@data_frame.summary.match("#{@data_frame.name}")).to_not eq(nil)
+      expect(@df_mi.summary.match("#{@df_mi.name}")).to_not eq(nil)
+    end
+  end
+  
+  context "#to_gsl" do
+    it "converts to GSL::Matrix" do
+      rows = [[1,2,3,4,5],[11,12,13,14,15],[11,22,33,44,55]].transpose
+      mat = GSL::Matrix.alloc *rows
+      expect(@data_frame.to_gsl).to eq(mat)
+    end
+  end
+
+  context "#merge" do
+    it "merges one dataframe with another" do
+      a = Daru::Vector.new [1, 2, 3]
+      b = Daru::Vector.new [3, 4, 5]
+      c = Daru::Vector.new [4, 5, 6]
+      d = Daru::Vector.new [7, 8, 9]
+      e = Daru::Vector.new [10, 20, 30]
+      ds1 = Daru::DataFrame.new({ :a => a, :b => b })
+      ds2 = Daru::DataFrame.new({ :c => c, :d => d })
+      exp = Daru::DataFrame.new({ :a => a, :b => b, :c => c, :d => d })
+
+      expect(ds1.merge(ds2)).to eq(exp)
+      expect(ds2.merge(ds1)).to eq(
+        Daru::DataFrame.new({c: c, d: d, a: a, b: b}, order: [:c, :d, :a, :b]))
+
+      ds3 = Daru::DataFrame.new({ :a => e })
+      exp = Daru::DataFrame.new({ :a_1 => a, :a_2 => e, :b => b }, 
+        order: [:a_1, :b, :a_2])
+
+      expect(ds1.merge(ds3)).to eq(exp)
+    end
+  end
+
+  context "#vector_by_calculation" do
+    it "DSL for returning vector of each calculation" do
+      a1 = Daru::Vector.new([1, 2, 3, 4, 5, 6, 7])
+      a2 = Daru::Vector.new([10, 20, 30, 40, 50, 60, 70])
+      a3 = Daru::Vector.new([100, 200, 300, 400, 500, 600, 700])
+      ds = Daru::DataFrame.new({ :a => a1, :b => a2, :c => a3 })
+      total = ds.vector_by_calculation { a + b + c }
+      expected = Daru::Vector.new([111, 222, 333, 444, 555, 666, 777])
+      expect(total).to eq(expected)
+    end
+  end
+
+  context "#vector_sum" do
+    before do
+      a1 = Daru::Vector.new [1, 2, 3, 4, 5, nil]
+      a2 = Daru::Vector.new [10, 10, 20, 20, 20, 30]
+      b1 = Daru::Vector.new [nil, 1, 1, 1, 1, 2]
+      b2 = Daru::Vector.new [2, 2, 2, nil, 2, 3]
+      @df = Daru::DataFrame.new({ :a1 => a1, :a2 => a2, :b1 => b1, :b2 => b2 })
+    end
+
+    it "calculates complete vector sum" do
+      expect(@df.vector_sum).to eq(Daru::Vector.new [nil, 15, 26, nil, 28, nil])
+    end
+
+    it "calculates partial vector sum" do
+      a = @df.vector_sum([:a1, :a2])
+      b = @df.vector_sum([:b1, :b2])
+
+      expect(a).to eq(Daru::Vector.new [11, 12, 23, 24, 25, nil])
+      expect(b).to eq(Daru::Vector.new [nil, 3, 3, nil, 3, 5])
+    end
+  end
+
+  context "#missing_values_rows" do
+    it "returns number of missing values in each row" do
+      a1 = Daru::Vector.new [1, nil, 3, 4, 5, nil]
+      a2 = Daru::Vector.new [10, nil, 20, 20, 20, 30]
+      b1 = Daru::Vector.new [nil, nil, 1, 1, 1, 2]
+      b2 = Daru::Vector.new [2, 2, 2, nil, 2, 3]
+      c  = Daru::Vector.new [nil, 2, 4, 2, 2, 2]
+      df = Daru::DataFrame.new({ 
+        :a1 => a1, :a2 => a2, :b1 => b1, :b2 => b2, :c => c })
+
+      expect(df.missing_values_rows).to eq(Daru::Vector.new [2, 3, 0, 1, 0, 1])
+    end
+  end
+
+  context "#vector_count_characters" do
+    it "" do
+      a1  = Daru::Vector.new( [1, 'abcde', 3, 4, 5, nil])
+      a2  = Daru::Vector.new( [10, 20.3, 20, 20, 20, 30])
+      b1  = Daru::Vector.new( [nil, '343434', 1, 1, 1, 2])
+      b2  = Daru::Vector.new( [2, 2, 2, nil, 2, 3])
+      c   = Daru::Vector.new([nil, 2, 'This is a nice example', 2, 2, 2])
+      ds  = Daru::DataFrame.new({ :a1 => a1, :a2 => a2, :b1 => b1, :b2 => b2, :c => c })
+
+      expect(ds.vector_count_characters).to eq(Daru::Vector.new([4, 17, 27, 5, 6, 5]))
+    end
+  end
+
+
+  context "has_missing_data?" do
+    before do
+      a1 = Daru::Vector.new [1, nil, 3, 4, 5, nil]
+      a2 = Daru::Vector.new [10, nil, 20, 20, 20, 30]
+      b1 = Daru::Vector.new [nil, nil, 1, 1, 1, 2]
+      b2 = Daru::Vector.new [2, 2, 2, nil, 2, 3]
+      c  = Daru::Vector.new [nil, 2, 4, 2, 2, 2]
+      @df = Daru::DataFrame.new({ :a1 => a1, :a2 => a2, :b1 => b1, :b2 => b2, :c => c })
+    end
+
+    it "returns true when missing data present" do
+      expect(@df.has_missing_data?).to eq(true)
+    end
+
+    it "returns false when no missing data prensent" do
+      a = @df.dup_only_valid
+      expect(a.has_missing_data?).to eq(false)
+    end
+  end
+
+  context "#vector_mean" do
+    before do
+      a1 = Daru::Vector.new [1, 2, 3, 4, 5, nil]
+      a2 = Daru::Vector.new [10, 10, 20, 20, 20, 30]
+      b1 = Daru::Vector.new [nil, 1, 1, 1, 1, 2]
+      b2 = Daru::Vector.new [2, 2, 2, nil, 2, 3]
+      c  = Daru::Vector.new [nil, 2, 4, 2, 2, 2]
+      @df = Daru::DataFrame.new({
+        :a1 => a1, :a2 => a2, :b1 => b1, :b2 => b2, :c => c })
+    end
+
+    it "calculates complete vector mean" do
+      expect(@df.vector_mean).to eq(
+        Daru::Vector.new [nil, 3.4, 6, nil, 6.0, nil])
+    end
+  end
+
+  context "#add_vectors_by_split_recode" do
+    before do
+      @ds = Daru::DataFrame.new({ 
+        :id   => Daru::Vector.new([1, 2, 3, 4, 5]), 
+        :name => Daru::Vector.new(%w(Alex Claude Peter Franz George)), 
+        :age  => Daru::Vector.new([20, 23, 25, 27, 5]),
+        :city => Daru::Vector.new(['New York', 'London', 'London', 'Paris', 'Tome']),
+        :a1   => Daru::Vector.new(['a,b', 'b,c', 'a', nil, 'a,b,c']) }, 
+        order: [:id, :name, :age, :city, :a1])
+    end
+
+    it "" do
+      @ds.add_vectors_by_split_recode(:a1, '_')
+      expect(@ds.vectors.to_a)    .to eq([:id, :name, :age, :city ,:a1, :a1_1, :a1_2, :a1_3])
+      expect(@ds[:a1_1].to_a).to eq([1, 0, 1, nil, 1])
+      expect(@ds[:a1_2].to_a).to eq([1, 1, 0, nil, 1])
+      expect(@ds[:a1_3].to_a).to eq([0, 1, 0, nil, 1])
+    end
+  end
+
+  context "#add_vectors_by_split" do
+    before do
+      @ds = Daru::DataFrame.new({ 
+        :id   => Daru::Vector.new([1, 2, 3, 4, 5]), 
+        :name => Daru::Vector.new(%w(Alex Claude Peter Franz George)), 
+        :age  => Daru::Vector.new([20, 23, 25, 27, 5]),
+        :city => Daru::Vector.new(['New York', 'London', 'London', 'Paris', 'Tome']),
+        :a1   => Daru::Vector.new(['a,b', 'b,c', 'a', nil, 'a,b,c']) 
+        }, order: [:id, :name, :age, :city, :a1])
+    end
+
+    it "" do
+      @ds.add_vectors_by_split(:a1, '_')
+      expect(@ds.vectors.to_a).to eq([:id, :name, :age, :city, :a1, :a1_a, :a1_b, :a1_c])
+      expect(@ds[:a1_a].to_a).to eq([1, 0, 1, nil, 1])
+      expect(@ds[:a1_b].to_a).to eq([1, 1, 0, nil, 1])
+      expect(@ds[:a1_c].to_a).to eq([0, 1, 0, nil, 1])
+    end
+  end
+
+  context "#verify" do
+    def create_test(*args, &_proc)
+      description = args.shift
+      fields = args
+      [description, fields, Proc.new]
+    end
+
+    before do
+      name = Daru::Vector.new %w(r1 r2 r3 r4)
+      v1   = Daru::Vector.new [1, 2, 3, 4]
+      v2   = Daru::Vector.new [4, 3, 2, 1]
+      v3   = Daru::Vector.new [10, 20, 30, 40]
+      v4   = Daru::Vector.new %w(a b a b)
+      @df = Daru::DataFrame.new({ 
+        :v1 => v1, :v2 => v2, :v3 => v3, :v4 => v4, :id => name 
+        }, order: [:v1, :v2, :v3, :v4, :id])
+    end
+
+    it "correctly verifies data as per the block" do
+      # Correct
+      t1 = create_test('If v4=a, v1 odd') do |r| 
+        r[:v4] == 'b' or (r[:v4] == 'a' and r[:v1].odd?)
+      end
+      t2 = create_test('v3=v1*10')  { |r| r[:v3] == r[:v1] * 10 }
+      # Fail!
+      t3 = create_test("v4='b'") { |r| r[:v4] == 'b' }
+      exp1 = ["1 [1]: v4='b'", "3 [3]: v4='b'"]
+      exp2 = ["1 [r1]: v4='b'", "3 [r3]: v4='b'"]
+
+      dataf = @df.verify(t3, t1, t2)
+      expect(dataf).to eq(exp1)
+
+      dataf = @df.verify(:id, t1, t2, t3)
+      expect(dataf).to eq(exp2)
+    end
+  end
+
+  context "#compute" do
+    it "performs a computation when supplied in a string" do
+      v1       = Daru::Vector.new [1, 2, 3, 4]
+      v2       = Daru::Vector.new [4, 3, 2, 1]
+      v3       = Daru::Vector.new [10, 20, 30, 40]
+      vnumeric = Daru::Vector.new [0, 0, 1, 4]
+      vsum     = Daru::Vector.new [1 + 4 + 10.0, 2 + 3 + 20.0, 3 + 2 + 30.0, 4 + 1 + 40.0]
+      vmult    = Daru::Vector.new [1 * 4, 2 * 3, 3 * 2, 4 * 1]
+
+      df = Daru::DataFrame.new({:v1 => v1, :v2 => v2, :v3 => v3})
+
+      expect(df.compute("v1/v2")).to eq(vnumeric)
+      expect(df.compute("v1+v2+v3")).to eq(vsum)
+      expect(df.compute("v1*v2")).to eq(vmult)
+    end
+  end
+
+  context ".crosstab_by_assignation" do
+    it "" do
+      v1 = Daru::Vector.new %w(a a a b b b c c c)
+      v2 = Daru::Vector.new %w(a b c a b c a b c)
+      v3 = Daru::Vector.new [0, 1, 0, 0, 1, 1, 0, 0, 1]
+      df = Daru::DataFrame.crosstab_by_assignation(v1, v2, v3)
+
+      expect(df[:_id].type).to eq(:object)
+      expect(df[:a].type).to eq(:numeric)
+      expect(df[:b].type).to eq(:numeric)
+
+      ev_id = Daru::Vector.new %w(a b c)
+      ev_a  = Daru::Vector.new [0, 0, 0]
+      ev_b  = Daru::Vector.new [1, 1, 0]
+      ev_c  = Daru::Vector.new [0, 1, 1]
+      df2 = Daru::DataFrame.new({ 
+        :_id => ev_id, :a => ev_a, :b => ev_b, :c => ev_c })
+
+      expect(df2).to eq(df)
+    end
+  end
+
+  context "#one_to_many" do
+    it "" do
+      rows = [
+        ['1', 'george', 'red', 10, 'blue', 20, nil, nil],
+        ['2', 'fred', 'green', 15, 'orange', 30, 'white', 20],
+        ['3', 'alfred', nil, nil, nil, nil, nil, nil]
+      ]
+      df = Daru::DataFrame.rows(rows, 
+        order: [:id, :name, :car_color1, :car_value1, :car_color2, 
+          :car_value2, :car_color3, :car_value3])
+
+      ids     = Daru::Vector.new %w(1 1 2 2 2)
+      colors  = Daru::Vector.new %w(red blue green orange white)
+      values  = Daru::Vector.new [10, 20, 15, 30, 20]
+      col_ids = Daru::Vector.new [1, 2, 1, 2, 3]
+      df_expected = Daru::DataFrame.new({
+        :id => ids, :_col_id => col_ids, :color => colors, :value => values
+        }, order: [:id, :_col_id, :color, :value])
+
+      expect(df.one_to_many([:id], 'car_%v%n')).to eq(df_expected)
+    end
+  end
+
+  context "#any?" do
+    before do
+      @df = Daru::DataFrame.new({
+        a: [1,2,3,4,5], 
+        b: [10,20,30,40,50],
+        c: [11,22,33,44,55]})
+    end
+
+    it "returns true if any one of the vectors satisfy condition" do
+      expect(@df.any? { |v| v[0] == 1 }).to eq(true)
+    end
+
+    it "returns false if none of the vectors satisfy the condition" do
+      expect(@df.any? { |v| v.mean > 100 }).to eq(false)
+    end
+
+    it "returns true if any one of the rows satisfy condition" do
+      expect(@df.any?(:row) { |r| r[:a] == 1 and r[:c] == 11 }).to eq(true)
+    end
+
+    it "returns false if none of the rows satisfy the condition" do
+      expect(@df.any?(:row) { |r| r.mean > 100 }).to eq(false)
+    end
+  end
+
+  context "#all?" do
+    before do
+      @df = Daru::DataFrame.new({
+        a: [1,2,3,4,5], 
+        b: [10,20,30,40,50],
+        c: [11,22,33,44,55]})
+    end
+    
+    it "returns true if all of the vectors satisfy condition" do
+      expect(@df.all? { |v| v.mean < 40 }).to eq(true)
+    end
+
+    it "returns false if any one of the vectors does not satisfy condition" do
+      expect(@df.all? { |v| v.mean == 30 }).to eq(false)
+    end
+
+    it "returns true if all of the rows satisfy condition" do
+      expect(@df.all?(:row) { |r| r.mean < 70 }).to eq(true)
+    end
+
+    it "returns false if any one of the rows does not satisfy condition" do
+      expect(@df.all?(:row) { |r| r.mean == 30 }).to eq(false)
+    end
+  end
+
+  context "#only_numerics" do
+    before do
+      @v1 = Daru::Vector.new([1,2,3,4,5])
+      @v2 = Daru::Vector.new(%w(one two three four five))
+      @v3 = Daru::Vector.new([11,22,33,44,55])
+      @df = Daru::DataFrame.new({
+        a: @v1, b: @v2, c: @v3 }, clone: false)
+    end
+
+    it "returns a view of only the numeric vectors" do
+      dfon = @df.only_numerics(clone: false)
+
+      expect(dfon).to eq(
+        Daru::DataFrame.new({ a: @v1, c: @v3 }, clone: false))
+      expect(dfon[:a].object_id).to eq(@v1.object_id)
+    end
+
+    it "returns a clone of numeric vectors" do
+      dfon = @df.only_numerics
+
+      expect(dfon).to eq(
+        Daru::DataFrame.new({ a: @v1, c: @v3}, clone: false)
+      )
+      expect(dfon[:a].object_id).to_not eq(@v1.object_id)
+    end
+
+    context Daru::MultiIndex do
+      before do
+        agg_vectors = Daru::MultiIndex.new(
+          [
+            [:d, :one, :large],
+            [:d, :one, :small],
+            [:d, :two, :large],
+            [:d, :two, :small],
+            [:e, :one, :large],
+            [:e, :one, :small],
+            [:e, :two, :large],
+            [:e, :two, :small]
+          ]
+        )
+
+        agg_index = Daru::MultiIndex.new(
+          [
+            [:bar],
+            [:foo]
+          ]
+        )
+        @df = Daru::DataFrame.new(
+          [
+            [4.112,2.234],
+            %w(a b),
+            [6.342,nil],
+            [7.2344,3.23214],
+            [8.234,4.533],
+            [10.342,2.3432],
+            [12.0,nil],
+            %w(a b)
+          ], order: agg_vectors, index: agg_index
+        )
+      end
+
+      it "returns numeric vectors" do
+        vectors = Daru::MultiIndex.new(
+          [
+            [:d, :one, :large],
+            [:d, :two, :large],
+            [:d, :two, :small],
+            [:e, :one, :large],
+            [:e, :one, :small],
+            [:e, :two, :large]
+          ]
+        )
+
+        index = Daru::MultiIndex.new(
+          [
+            [:bar],
+            [:foo]
+          ]
+        )
+        answer = Daru::DataFrame.new(
+          [
+            [4.112,2.234],
+            [6.342,nil],
+            [7.2344,3.23214],
+            [8.234,4.533],
+            [10.342,2.3432],
+            [12.0,nil],
+          ], order: vectors, index: index
+        )
+
+        expect(@df.only_numerics).to eq(answer)
+      end
+    end
+  end
 end if mri?