回答編集履歴
2
追記
answer
CHANGED
|
@@ -8,7 +8,7 @@
|
|
|
8
8
|
```
|
|
9
9
|
|
|
10
10
|
コードの速さが気になるのであれば、numpy.arrayへのキャストなくすべきでしょう。
|
|
11
|
-
|
|
11
|
+
また検索があるので、listやnp.arrayではなくsetを使うとだいぶ早くなります。
|
|
12
12
|
```python
|
|
13
13
|
import math
|
|
14
14
|
import pandas as pd
|
|
@@ -16,7 +16,7 @@
|
|
|
16
16
|
grundy_ceil = [0]
|
|
17
17
|
|
|
18
18
|
for n in range(1, 101):
|
|
19
|
-
a = [grundy_ceil[n - i] for i in range(1, math.ceil(n /2 + 1))]
|
|
19
|
+
a = set([grundy_ceil[n - i] for i in range(1, math.ceil(n /2 + 1))])
|
|
20
20
|
grundy_ceil.append(next(i for i in count() if i not in a))
|
|
21
21
|
|
|
22
22
|
df = pd.DataFrame(grundy_ceil).T
|
|
@@ -24,6 +24,9 @@
|
|
|
24
24
|
|
|
25
25
|
---
|
|
26
26
|
追記：完全に趣味ですが。[ここ](https://stackoverflow.com/questions/7088625/what-is-the-most-efficient-way-to-check-if-a-value-exists-in-a-numpy-array)を参考に。
|
|
27
|
+
yag1kazさんのlistからの取り出し方も取り入れました。
|
|
28
|
+
sliceをリスト内包表記で書くと遅くなるのは盲点でした…
|
|
29
|
+
|
|
27
30
|
```python
|
|
28
31
|
from timeit import timeit
|
|
29
32
|
from itertools import count
|
|
@@ -74,14 +77,42 @@
|
|
|
74
77
|
df = pd.DataFrame(grundy_ceil).T
|
|
75
78
|
return df
|
|
76
79
|
|
|
80
|
+
def f4():
|
|
81
|
+
grundy_ceil = [0]
|
|
82
|
+
for n in range(1, 101):
|
|
83
|
+
a = set(grundy_ceil[:-math.ceil(n/2+1):-1])
|
|
84
|
+
j = next(i for i in count() if i not in a)
|
|
85
|
+
grundy_ceil.append(j)
|
|
86
|
+
df = pd.DataFrame(grundy_ceil).T
|
|
87
|
+
return df
|
|
88
|
+
|
|
89
|
+
a = f0()
|
|
90
|
+
b = f1()
|
|
91
|
+
c = f2()
|
|
92
|
+
d = f3()
|
|
93
|
+
e = f4()
|
|
94
|
+
print('f1', np.allclose(a[0].values, b[0].values))
|
|
95
|
+
print('f2', np.allclose(a[0].values, c[0].values))
|
|
96
|
+
print('f3', np.allclose(a[0].values, d[0].values))
|
|
97
|
+
print('f4', np.allclose(a[0].values, e[0].values))
|
|
98
|
+
'''
|
|
99
|
+
一応同じ結果になることを確認
|
|
100
|
+
f1 True
|
|
101
|
+
f2 True
|
|
102
|
+
f3 True
|
|
103
|
+
f4 True
|
|
104
|
+
'''
|
|
77
105
|
n = 1000
|
|
78
|
-
print('Original
|
|
106
|
+
print('Original {}'.format(timeit(f0, number=n)))
|
|
79
|
-
print('list
|
|
107
|
+
print('list {}'.format(timeit(f1, number=n)))
|
|
80
|
-
print('np.array
|
|
108
|
+
print('np.array {}'.format(timeit(f2, number=n)))
|
|
81
|
-
print('set
|
|
109
|
+
print('set {}'.format(timeit(f3, number=n)))
|
|
82
|
-
|
|
110
|
+
print('slice+set {}'.format(timeit(f4, number=n)))
|
|
111
|
+
'''
|
|
83
|
-
|
|
112
|
+
Original 9.65767323700129
|
|
84
|
-
|
|
113
|
+
list 1.1692692330107093
|
|
85
|
-
|
|
114
|
+
np.array 7.38920716199209
|
|
86
|
-
|
|
115
|
+
set 0.9272018460032996
|
|
116
|
+
slice+set 0.6420557640085462
|
|
117
|
+
'''
|
|
87
118
|
```
|
1
追記
answer
CHANGED
|
@@ -20,4 +20,68 @@
|
|
|
20
20
|
grundy_ceil.append(next(i for i in count() if i not in a))
|
|
21
21
|
|
|
22
22
|
df = pd.DataFrame(grundy_ceil).T
|
|
23
|
+
```
|
|
24
|
+
|
|
25
|
+
---
|
|
26
|
+
追記：完全に趣味ですが。[ここ](https://stackoverflow.com/questions/7088625/what-is-the-most-efficient-way-to-check-if-a-value-exists-in-a-numpy-array)を参考に。
|
|
27
|
+
```python
|
|
28
|
+
from timeit import timeit
|
|
29
|
+
from itertools import count
|
|
30
|
+
import math
|
|
31
|
+
import pandas as pd
|
|
32
|
+
import numpy as np
|
|
33
|
+
|
|
34
|
+
pd.options.display.max_columns = None
|
|
35
|
+
pd.options.display.notebook_repr_html = True
|
|
36
|
+
|
|
37
|
+
def f0():
|
|
38
|
+
grundy_ceil = np.array([0])
|
|
39
|
+
#grundy_ceil_1p = np.array([0])
|
|
40
|
+
for n in range(1, 101):
|
|
41
|
+
compare = np.array([])
|
|
42
|
+
compare = np.append(compare, [grundy_ceil[n - i] for i in range(1, math.ceil(n /2 + 1))])
|
|
43
|
+
for j in range(compare.size + 1):
|
|
44
|
+
if not j in compare:
|
|
45
|
+
grundy_ceil = np.append(grundy_ceil, j)
|
|
46
|
+
break
|
|
47
|
+
df = pd.DataFrame(grundy_ceil).T
|
|
48
|
+
return df
|
|
49
|
+
|
|
50
|
+
def f1():
|
|
51
|
+
grundy_ceil = [0]
|
|
52
|
+
for n in range(1, 101):
|
|
53
|
+
a = [grundy_ceil[n - i] for i in range(1, math.ceil(n /2 + 1))]
|
|
54
|
+
j = next(i for i in count() if i not in a)
|
|
55
|
+
grundy_ceil.append(j)
|
|
56
|
+
df = pd.DataFrame(grundy_ceil).T
|
|
57
|
+
return df
|
|
58
|
+
|
|
59
|
+
def f2():
|
|
60
|
+
grundy_ceil = [0]
|
|
61
|
+
for n in range(1, 101):
|
|
62
|
+
a = np.array([grundy_ceil[n - i] for i in range(1, math.ceil(n /2 + 1))])
|
|
63
|
+
j = next(i for i in count() if i not in a)
|
|
64
|
+
grundy_ceil.append(j)
|
|
65
|
+
df = pd.DataFrame(grundy_ceil).T
|
|
66
|
+
return df
|
|
67
|
+
|
|
68
|
+
def f3():
|
|
69
|
+
grundy_ceil = [0]
|
|
70
|
+
for n in range(1, 101):
|
|
71
|
+
a = set([grundy_ceil[n - i] for i in range(1, math.ceil(n /2 + 1))])
|
|
72
|
+
j = next(i for i in count() if i not in a)
|
|
73
|
+
grundy_ceil.append(j)
|
|
74
|
+
df = pd.DataFrame(grundy_ceil).T
|
|
75
|
+
return df
|
|
76
|
+
|
|
77
|
+
n = 1000
|
|
78
|
+
print('Original {}'.format(timeit(f0, number=n)))
|
|
79
|
+
print('list {}'.format(timeit(f1, number=n)))
|
|
80
|
+
print('np.array {}'.format(timeit(f2, number=n)))
|
|
81
|
+
print('set {}'.format(timeit(f3, number=n)))
|
|
82
|
+
|
|
83
|
+
# Original 9.262939481996
|
|
84
|
+
# list 1.147810668015154
|
|
85
|
+
# np.array 7.4433728019939736
|
|
86
|
+
# set 0.8586325150099583
|
|
23
87
|
```
|