回答編集履歴
2
補足を追加
test
CHANGED
|
@@ -91,3 +91,45 @@
|
|
|
91
91
|
limit -= drop_size
|
|
92
92
|
|
|
93
93
|
```
|
|
94
|
+
|
|
95
|
+
|
|
96
|
+
|
|
97
|
+
---
|
|
98
|
+
|
|
99
|
+
**【追記２】**
|
|
100
|
+
|
|
101
|
+
だいたい、こんな感じになるのではないでしょうか（動作は全く未検証）
|
|
102
|
+
|
|
103
|
+
```Python
|
|
104
|
+
|
|
105
|
+
import pandas as pd
|
|
106
|
+
|
|
107
|
+
import numpy as np
|
|
108
|
+
|
|
109
|
+
|
|
110
|
+
|
|
111
|
+
for list_no in range(1,3):
|
|
112
|
+
|
|
113
|
+
df = pd.read_csv(f"list{list_no}.csv")
|
|
114
|
+
|
|
115
|
+
label = pd.cut(df['time2'], bins=np.arange(0, 10.5, 0.5), right=False, labels=False)
|
|
116
|
+
|
|
117
|
+
|
|
118
|
+
|
|
119
|
+
limit = 5
|
|
120
|
+
|
|
121
|
+
for i in range(0,20):
|
|
122
|
+
|
|
123
|
+
drop_size = min((label == i).sum(), limit)
|
|
124
|
+
|
|
125
|
+
drop_index = df.loc[label==i,'time2'].sort_values().index[:drop_size]
|
|
126
|
+
|
|
127
|
+
df = df.drop(drop_index)
|
|
128
|
+
|
|
129
|
+
new_file = f"list{list_no}_{(i+1)*0.5:.1f}f.csv"
|
|
130
|
+
|
|
131
|
+
df.to_csv(new_file)
|
|
132
|
+
|
|
133
|
+
limit -= drop_size
|
|
134
|
+
|
|
135
|
+
```
|
1
サンプルを追加
test
CHANGED
|
@@ -65,3 +65,29 @@
|
|
|
65
65
|
```
|
|
66
66
|
|
|
67
67
|
のようにtime2の値でソートして頭からn個のIndex値を対象とすると良いのではないでしょうか。
|
|
68
|
+
|
|
69
|
+
|
|
70
|
+
|
|
71
|
+
---
|
|
72
|
+
|
|
73
|
+
**追記**
|
|
74
|
+
|
|
75
|
+
削除する行数を全体で制限する場合のサンプル
|
|
76
|
+
|
|
77
|
+
```
|
|
78
|
+
|
|
79
|
+
limit = 5
|
|
80
|
+
|
|
81
|
+
for i in range(0,20):
|
|
82
|
+
|
|
83
|
+
drop_size = min((label == i).sum(), limit)
|
|
84
|
+
|
|
85
|
+
drop_index = df.loc[label==i,'time2'].sort_values().index[:drop_size]
|
|
86
|
+
|
|
87
|
+
df = df.drop(drop_index)
|
|
88
|
+
|
|
89
|
+
df.to_csv(f"list1_{(i+1)*0.5:.1f}f.csv")
|
|
90
|
+
|
|
91
|
+
limit -= drop_size
|
|
92
|
+
|
|
93
|
+
```
|