問題解決：SettingWithCopyWarning: A value is trying to be set on a copy of a slice from a DataFrame

轉載：y小川

SettingWithCopyWarning 解決方案

問題場景：我在讀取csv文件之后，因為要新增一個特征列並根據已有特征修改新增列的值，結果在修改的時候就碰到了SettingWithCopyWarning這個警告，花了很長時間才解決這個問題。

一個簡易版的范例

import pandas as pd
import numpy as np

aa = np.array([1, 0, 1, 0])
bb = pd.DataFrame(aa.T, columns=['one'])
print(bb)

輸出為：

添加一個新列后在輸出

bb['two'] = 0
print(bb)

output[]:
   one  two
0    1    0
1    0    0
2    1    0
3    0    0

按條件修改新列再輸出就報錯了：

for i in range(bb.shape[0]):
    if bb['one'][i] == 0:
        bb['two'][i] = 1
print(bb)

output[]:
C:/PycharmProjects/NaiveBayesProduct/pandas/try_index.py:22: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame

See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy
  bb['two'][i] = 1
   one  two
0    1    0
1    0    1
2    1    0
3    0    1

這個問題怎么解決呢，我查了stackoverflow上的很多帖子，試了loc/iloc等函數都不管用，最后才發現是順序錯了。正確方案應該是生成好正確的數組再插入dataframe中。下面我把上面的例子用正確地方法再重新生成一遍。

import pandas as pd
import numpy as np

aa = np.array([1, 0, 1, 0])
bb = pd.DataFrame(aa.T, columns=['one'])
# 生成一個ndarray，裝要插入的值
two = np.zeros(bb.shape[0])
# 按條件修改two
for i in range(bb.shape[0]):
    if bb['one'][i] == 0:
        two[i] = 1
# 完成后將two插入dataframe中
bb.insert(1,'two', two)
print(bb)

output[]:
   one  two
0    1  0.0
1    0  1.0
2    1  0.0
3    0  1.0