Module: Daru::Core::MergeHelper

Defined in:
lib/daru/core/merge.rb

Class Method Summary collapse

Class Method Details

.arrayify(df) ⇒ Object 



37
38
39
40
41
42
43
 
# File 'lib/daru/core/merge.rb', line 37

def arrayify df
  arr = df.to_a
  col_names = arr[0][0].keys
  values = arr[0].map{|h| h.values}

  return col_names, values
end

.bf_inner_join(df1, df2, on) ⇒ Object 



64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
 
# File 'lib/daru/core/merge.rb', line 64

def bf_inner_join df1, df2, on
  col_names1, table1 = arrayify df1
  col_names2, table2 = arrayify df2

  #resolve duplicates
  indicies1 = on.map{|i| col_names1.index(i)}
  indicies2 = on.map{|i| col_names2.index(i)}
  col_names2.map! do |name| 
    if (col_names1.include?(name))
      col_names1[col_names1.index(name)] = (name.to_s + "_1").to_sym unless on.include?(name)
      (name.to_s + "_2").to_sym
    else
      name
    end
  end

  #combine key columns to a single column value
  on_cols1 = table1.flat_map{|x| indicies1.map{|i| x[i].to_s}.join("+")}
  on_cols2 = table2.flat_map{|x| indicies2.map{|i| x[i].to_s}.join("+")}

  #parameters for a BF with approx 0.1% false positives
  m = on_cols2.size * 15
  k = 11

  bf = BloomFilter::Native.new({:size => m, :hashes => k, :bucket => 1})
  on_cols2.each{|x| bf.insert(x)}

  x_ind = -1
  joined_new = on_cols1.map do |x|
    x_ind+=1
    if (bf.include?(x))
      {x_ind => on_cols2.each_index.select{|y_ind| on_cols2[y_ind] == x}}
    else
      {x_ind => []}
    end
  end
    .reduce({}) {|h,pairs| pairs.each {|k,v| (h[k] ||= []) << v}; h}
    .flat_map{|ind1, inds2| inds2.flatten.map{|ind2| [table1[ind1], table2[ind2]].flatten} if inds2.flatten.size > 0}

  joined_cols = [col_names1, col_names2].flatten
  df = Daru::DataFrame.rows(joined_new.compact, order: joined_cols)
  on.each{|x| df.delete_vector (x.to_s + "_2").to_sym}

  df
end

.full_outer_join(df1, df2, df_hash1, df_hash2, on) ⇒ Object 



110
111
112
113
114
115
116
 
# File 'lib/daru/core/merge.rb', line 110

def full_outer_join df1, df2, df_hash1, df_hash2, on
  left  = left_outer_join df1, df2, df_hash1, df_hash2, on, true
  right = right_outer_join df1, df2, df_hash1, df_hash2, on, true

  Daru::DataFrame.rows(
    (left.values.transpose | right.values.transpose), order: left.keys)
end

.hashify(df) ⇒ Object 



31
32
33
34
35
 
# File 'lib/daru/core/merge.rb', line 31

def hashify df
  hsh = df.to_hash
  hsh.each { |k,v| hsh[k] = v.to_a }
  hsh
end

.inner_join(df1, df2, df_hash1, df_hash2, on) ⇒ Object 



45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
 
# File 'lib/daru/core/merge.rb', line 45

def inner_join df1, df2, df_hash1, df_hash2, on
  joined_hash = {}
  ((df_hash1.keys - on) | on | (df_hash2.keys - on)).each do |k|
    joined_hash[k] = []
  end

  (0...df1.size).each do |id1|
    (0...df2.size).each do |id2|
      if on.all? { |n| df_hash1[n][id1] == df_hash2[n][id2] }
        joined_hash.each do |k,v|
          v << (df_hash1.has_key?(k) ? df_hash1[k][id1] : df_hash2[k][id2])
        end
      end
    end
  end

  Daru::DataFrame.new(joined_hash, order: joined_hash.keys)
end

.left_outer_join(df1, df2, df_hash1, df_hash2, on, as_hash = false) ⇒ Object 



118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
 
# File 'lib/daru/core/merge.rb', line 118

def left_outer_join df1, df2, df_hash1, df_hash2, on, as_hash=false
  joined_hash = {}
  ((df_hash1.keys - on) | on | (df_hash2.keys - on)).each do |k|
    joined_hash[k] = []
  end

  
  (0...df1.size).each do |id1|
    joined = false
    (0...df2.size).each do |id2|
      if on.all? { |n| df_hash1[n][id1] == df_hash2[n][id2] }
        joined = true
        joined_hash.each do |k,v|
          v << (df_hash1.has_key?(k) ? df_hash1[k][id1] : df_hash2[k][id2])
        end
      end
    end

    unless joined
      df_hash1.keys.each do |k|
        joined_hash[k] << df_hash1[k][id1]
      end

      (joined_hash.keys - df_hash1.keys).each do |k|
        joined_hash[k] << nil
      end
      joined = false
    end
  end

  return joined_hash if as_hash
  Daru::DataFrame.new(joined_hash, order: joined_hash.keys)
end

.replace_keys_if_duplicates(hash, matcher) ⇒ Object 



5
6
7
8
9
10
11
12
13
14
15
16
17
18
 
# File 'lib/daru/core/merge.rb', line 5

def replace_keys_if_duplicates hash, matcher
  matched = nil
  hash.keys.each { |d| 
    if matcher.match(Regexp.new(d.to_s))
      matched = d
      break
    end  
  }

  if matched
    hash[matcher] = hash[matched] 
    hash.delete matched
  end
end

.resolve_duplicates(df_hash1, df_hash2, on) ⇒ Object 



20
21
22
23
24
25
26
27
28
29
 
# File 'lib/daru/core/merge.rb', line 20

def resolve_duplicates df_hash1, df_hash2, on
  hk = df_hash1.keys + df_hash2.keys - on
  recoded = hk.recode_repeated.map(&:to_sym)
  diff = (recoded - hk).sort

  diff.each_slice(2) do |a|
    replace_keys_if_duplicates df_hash1, a[0]
    replace_keys_if_duplicates df_hash2, a[1]
  end
end

.right_outer_join(df1, df2, df_hash1, df_hash2, on, as_hash = false) ⇒ Object 



152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
 
# File 'lib/daru/core/merge.rb', line 152

def right_outer_join df1, df2, df_hash1, df_hash2, on, as_hash=false
  joined_hash = {}
  ((df_hash1.keys - on) | on | (df_hash2.keys - on)).each do |k|
    joined_hash[k] = []
  end

  (0...df2.size).each do |id1|
    joined = false
    (0...df1.size).each do |id2|
      if on.all? { |n| df_hash2[n][id1] == df_hash1[n][id2] }
        joined = true
        joined_hash.each do |k,v|
          v << (df_hash2.has_key?(k) ? df_hash2[k][id1] : df_hash1[k][id2])
        end
      end
    end

    unless joined
      df_hash2.keys.each do |k|
        joined_hash[k] << df_hash2[k][id1]
      end

      (joined_hash.keys - df_hash2.keys).each do |k|
        joined_hash[k] << nil
      end
      joined = false
    end
  end

  return joined_hash if as_hash
  Daru::DataFrame.new(joined_hash, order: joined_hash.keys)
end

.verify_dataframes(df_hash1, df_hash2, on) ⇒ Object

Raises:

  • (ArgumentError) 


185
186
187
188
189
190
 
# File 'lib/daru/core/merge.rb', line 185

def verify_dataframes df_hash1, df_hash2, on
  raise ArgumentError, 
    "All fields in :on must be present in self" if !on.all? { |e| df_hash1[e] }
  raise ArgumentError,
    "All fields in :on must be present in other DF" if !on.all? { |e| df_hash2[e] }
end