回答編集履歴
1
コード修正
test
CHANGED
|
@@ -8,7 +8,7 @@
|
|
|
8
8
|
|
|
9
9
|
|
|
10
10
|
|
|
11
|
-
df1 = pd.DataFrame({'name': ['A','B','C','D','E'],'a':[1,5,5,8,9]})
|
|
11
|
+
df1 = pd.DataFrame({'name': ['A','B','C','D','D-2','E'],'a':[1,5,5,5,8,9]})
|
|
12
12
|
|
|
13
13
|
df2 = pd.DataFrame({'table': ['2','3-1','3-2','7'],'a':[2,3,3,7]})
|
|
14
14
|
|
|
@@ -16,31 +16,47 @@
|
|
|
16
16
|
|
|
17
17
|
# df2の列値をタプル('table','a')のリストに展開
|
|
18
18
|
|
|
19
|
-
|
|
19
|
+
lst2 = [(r[0],r[1]) for r in df2.values]
|
|
20
20
|
|
|
21
21
|
|
|
22
22
|
|
|
23
|
-
# df2から指定値に最も近い
|
|
23
|
+
# df2から指定値に最も近いlst2の要素位置リストを返す
|
|
24
24
|
|
|
25
25
|
def nearest(a):
|
|
26
26
|
|
|
27
|
-
i
|
|
27
|
+
m = min(lst2, key=lambda v:(v[1]-a)*(v[1]-a))
|
|
28
28
|
|
|
29
|
-
return np.nan
|
|
30
|
-
|
|
31
|
-
|
|
29
|
+
return [i for i,v in enumerate(lst2) if v[1] == m[1]]
|
|
32
|
-
|
|
33
|
-
del df2_list[df2_list.index(m)] # df1に紐づけ済みは削除
|
|
34
|
-
|
|
35
|
-
return m[0]
|
|
36
30
|
|
|
37
31
|
|
|
38
32
|
|
|
33
|
+
used = set() # df1に割当済のlst2の要素位置
|
|
34
|
+
|
|
35
|
+
|
|
36
|
+
|
|
39
|
-
# 割り当て
|
|
37
|
+
# 最近傍値に割り当て
|
|
40
38
|
|
|
41
39
|
for i,r in df1.iterrows():
|
|
42
40
|
|
|
41
|
+
val = np.nan
|
|
42
|
+
|
|
43
|
+
idxs = nearest(r['a'])
|
|
44
|
+
|
|
45
|
+
while idxs:
|
|
46
|
+
|
|
47
|
+
idx = idxs.pop(0)
|
|
48
|
+
|
|
49
|
+
if idx not in used: # 未割当
|
|
50
|
+
|
|
51
|
+
val = lst2[idx][0]
|
|
52
|
+
|
|
53
|
+
used.add(idx) # 割当済を保持
|
|
54
|
+
|
|
55
|
+
break
|
|
56
|
+
|
|
57
|
+
|
|
58
|
+
|
|
43
|
-
df1.loc[i,'table'] =
|
|
59
|
+
df1.loc[i,'table'] = val
|
|
44
60
|
|
|
45
61
|
|
|
46
62
|
|
|
@@ -56,9 +72,11 @@
|
|
|
56
72
|
|
|
57
73
|
2 C 5 3-2
|
|
58
74
|
|
|
59
|
-
3 D
|
|
75
|
+
3 D 5 NaN
|
|
60
76
|
|
|
77
|
+
4 D-2 8 7
|
|
78
|
+
|
|
61
|
-
|
|
79
|
+
5 E 9 NaN
|
|
62
80
|
|
|
63
81
|
"""
|
|
64
82
|
|