回答編集履歴
2
挿入ソート版追記
answer
CHANGED
|
@@ -144,4 +144,83 @@
|
|
|
144
144
|
[ 3. 6. nan nan nan]
|
|
145
145
|
[ 4. 7. nan nan nan]]
|
|
146
146
|
"""
|
|
147
|
+
```
|
|
148
|
+
挿入ソート版
|
|
149
|
+
-----
|
|
150
|
+
題意から以下の処理でもよさそうです。
|
|
151
|
+
再帰版よりもはるかに速く処理できます。
|
|
152
|
+
|
|
153
|
+
```Python
|
|
154
|
+
import numpy as np
|
|
155
|
+
import pprint
|
|
156
|
+
|
|
157
|
+
import pandas as pd
|
|
158
|
+
from io import StringIO
|
|
159
|
+
f = StringIO("""c1,c2,c3,c4,c5
|
|
160
|
+
,,7,8,
|
|
161
|
+
1,2,,,
|
|
162
|
+
,,6,6,
|
|
163
|
+
4,7,,,
|
|
164
|
+
,,,2,1
|
|
165
|
+
,,,7,4
|
|
166
|
+
6,9,,,
|
|
167
|
+
,1,2,,
|
|
168
|
+
,,,9,5
|
|
169
|
+
,4,4,,
|
|
170
|
+
,,1,1,
|
|
171
|
+
2,3,,,
|
|
172
|
+
5,8,,,
|
|
173
|
+
,,,3,2
|
|
174
|
+
,,3,4,
|
|
175
|
+
,5,5,,
|
|
176
|
+
,,,5,3
|
|
177
|
+
3,6,,,""")
|
|
178
|
+
ary = pd.read_csv(f).values.tolist()
|
|
179
|
+
|
|
180
|
+
|
|
181
|
+
ret = []
|
|
182
|
+
|
|
183
|
+
# 各列について左から順に処理
|
|
184
|
+
col_cnt = len(ary[0])
|
|
185
|
+
for c in range(col_cnt):
|
|
186
|
+
|
|
187
|
+
# 対象列がnanでない行のみ抜き出す
|
|
188
|
+
rows = []
|
|
189
|
+
for r in ary[::-1]:
|
|
190
|
+
if not np.isnan(r[c]):
|
|
191
|
+
rows.append(r)
|
|
192
|
+
ary.remove(r)
|
|
193
|
+
|
|
194
|
+
# 結果配列に列値が昇順になるように挿入していく
|
|
195
|
+
for row in rows:
|
|
196
|
+
is_ins = False
|
|
197
|
+
for idx,ret_row in enumerate(ret):
|
|
198
|
+
if row[c] < ret_row[c]:
|
|
199
|
+
ret.insert(idx,row)
|
|
200
|
+
is_ins = True
|
|
201
|
+
break
|
|
202
|
+
if not is_ins:
|
|
203
|
+
ret.append(row)
|
|
204
|
+
|
|
205
|
+
pprint.pprint(ret)
|
|
206
|
+
"""
|
|
207
|
+
[[nan, nan, 1.0, 1.0, nan],
|
|
208
|
+
[nan, 1.0, 2.0, nan, nan],
|
|
209
|
+
[1.0, 2.0, nan, nan, nan],
|
|
210
|
+
[2.0, 3.0, nan, nan, nan],
|
|
211
|
+
[nan, nan, nan, 2.0, 1.0],
|
|
212
|
+
[nan, nan, nan, 3.0, 2.0],
|
|
213
|
+
[nan, nan, 3.0, 4.0, nan],
|
|
214
|
+
[nan, 4.0, 4.0, nan, nan],
|
|
215
|
+
[nan, 5.0, 5.0, nan, nan],
|
|
216
|
+
[3.0, 6.0, nan, nan, nan],
|
|
217
|
+
[4.0, 7.0, nan, nan, nan],
|
|
218
|
+
[5.0, 8.0, nan, nan, nan],
|
|
219
|
+
[6.0, 9.0, nan, nan, nan],
|
|
220
|
+
[nan, nan, nan, 5.0, 3.0],
|
|
221
|
+
[nan, nan, 6.0, 6.0, nan],
|
|
222
|
+
[nan, nan, nan, 7.0, 4.0],
|
|
223
|
+
[nan, nan, 7.0, 8.0, nan],
|
|
224
|
+
[nan, nan, nan, 9.0, 5.0]]
|
|
225
|
+
"""
|
|
147
226
|
```
|
1
再帰版を追記
answer
CHANGED
|
@@ -57,4 +57,91 @@
|
|
|
57
57
|
[nan 4. 4. nan nan]
|
|
58
58
|
[nan nan nan 3. 2.]]
|
|
59
59
|
"""
|
|
60
|
+
```
|
|
61
|
+
|
|
62
|
+
|
|
63
|
+
再帰版
|
|
64
|
+
-----
|
|
65
|
+
|
|
66
|
+
総当たりよりは速いですが、15行程度が限界ですね。
|
|
67
|
+
```Python
|
|
68
|
+
|
|
69
|
+
import numpy as np
|
|
70
|
+
|
|
71
|
+
def search( ary):
|
|
72
|
+
row_cnt = ary.shape[0]
|
|
73
|
+
col_cnt = ary.shape[1]
|
|
74
|
+
|
|
75
|
+
# 条件を満たすか
|
|
76
|
+
# row : 行の位置
|
|
77
|
+
# mins : 現時点の各列の最小値
|
|
78
|
+
def is_match(row,mins):
|
|
79
|
+
for col in range(col_cnt):
|
|
80
|
+
v = ary[row,col]
|
|
81
|
+
if np.isnan(v):
|
|
82
|
+
continue
|
|
83
|
+
if v < mins[col]:
|
|
84
|
+
return False
|
|
85
|
+
mins[col] = v # 最小値を更新
|
|
86
|
+
return True
|
|
87
|
+
|
|
88
|
+
# rows : 行位置の配列
|
|
89
|
+
# mins : 現時点の各列の最小値
|
|
90
|
+
def search_row(rows,mins):
|
|
91
|
+
if len(rows) == row_cnt:
|
|
92
|
+
return rows
|
|
93
|
+
|
|
94
|
+
rows_set = set(rows)
|
|
95
|
+
for row in range(row_cnt):
|
|
96
|
+
if row in rows_set: # 重複は除く
|
|
97
|
+
continue
|
|
98
|
+
next_mins = mins.copy()
|
|
99
|
+
if is_match(row,next_mins):
|
|
100
|
+
ret = search_row(rows+[row],next_mins)
|
|
101
|
+
if ret:
|
|
102
|
+
return ret
|
|
103
|
+
|
|
104
|
+
rows = search_row([],np.zeros(col_cnt))
|
|
105
|
+
return ary[rows,:]
|
|
106
|
+
|
|
107
|
+
|
|
108
|
+
import pandas as pd
|
|
109
|
+
from io import StringIO
|
|
110
|
+
f = StringIO("""c1,c2,c3,c4,c5
|
|
111
|
+
,,7,8,
|
|
112
|
+
1,2,,,
|
|
113
|
+
,,6,6,
|
|
114
|
+
4,7,,,
|
|
115
|
+
,,,2,1
|
|
116
|
+
,,,7,4
|
|
117
|
+
,1,2,,
|
|
118
|
+
,4,4,,
|
|
119
|
+
,,1,1,
|
|
120
|
+
2,3,,,
|
|
121
|
+
,,,3,2
|
|
122
|
+
,,3,4,
|
|
123
|
+
,5,5,,
|
|
124
|
+
,,,5,3
|
|
125
|
+
3,6,,,""")
|
|
126
|
+
ary = pd.read_csv(f).values
|
|
127
|
+
|
|
128
|
+
ret = search(ary)
|
|
129
|
+
print(ret)
|
|
130
|
+
"""
|
|
131
|
+
[[nan nan 1. 1. nan]
|
|
132
|
+
[nan nan nan 2. 1.]
|
|
133
|
+
[nan 1. 2. nan nan]
|
|
134
|
+
[ 1. 2. nan nan nan]
|
|
135
|
+
[ 2. 3. nan nan nan]
|
|
136
|
+
[nan nan nan 3. 2.]
|
|
137
|
+
[nan nan 3. 4. nan]
|
|
138
|
+
[nan 4. 4. nan nan]
|
|
139
|
+
[nan 5. 5. nan nan]
|
|
140
|
+
[nan nan nan 5. 3.]
|
|
141
|
+
[nan nan 6. 6. nan]
|
|
142
|
+
[nan nan nan 7. 4.]
|
|
143
|
+
[nan nan 7. 8. nan]
|
|
144
|
+
[ 3. 6. nan nan nan]
|
|
145
|
+
[ 4. 7. nan nan nan]]
|
|
146
|
+
"""
|
|
60
147
|
```
|