我從一個(gè)數(shù)據(jù)幀開始，我想根據(jù)輸入值的條件將它分成三個(gè)數(shù)據(jù)幀。對(duì)于 df1，我只想讓 Person_X 值 <0.33。對(duì)于 df2，我希望只有 0.34<Person_X 值<0.65。對(duì)于 df3，我只想讓 Person_X 值 >0.66。我希望 NaN 替換不符合要求的整數(shù)值。所以，從 df 開始：In [1]: df=pd.DataFrame({'location':[4991, 8870, 2703, 9674], ...: 'Person_1': ['NaN', 0.2,0.5,0.7], ...: 'Person_2':[0.8, 0.45, 'NaN', 0.1]}) ...: df Out[1]: location Person_1 Person_20 4991 NaN 0.81 8870 0.2 0.452 2703 0.5 NaN3 9674 0.7 0.1我想要三個(gè)看起來(lái)像的數(shù)據(jù)框： df1 location Person_1 Person_20 4991 NaN NaN1 8870 0.2 NaN2 2703 NaN NaN3 9674 NaN 0.1df2 Out[4]: location Person_1 Person_20 4991 NaN NaN1 8870 NaN 0.452 2703 0.5 NaN3 9674 NaN NaNdf3 Out[4]: location Person_1 Person_20 4991 NaN 0.81 8870 NaN NaN2 2703 NaN NaN3 9674 0.7 NaN我已經(jīng)針對(duì)每個(gè)要求嘗試過類似的方法，但是由于 df 已經(jīng)包含 NaN，所以我遇到了問題。patient_cols=[col for col in df if col.startswith('Person')]df[patient_cols]=df[patient_cols].applymap(lambda x: np.nan if x>0.33 else x)df[patient_cols]
1 回答
烙印99
TA貢獻(xiàn)1829條經(jīng)驗(yàn) 獲得超13個(gè)贊
讓我們使用pd.DataFrame.where：
df=pd.DataFrame({'location':[4991, 8870, 2703, 9674],
'Person_1': [np.nan, 0.2,0.5,0.7],
'Person_2':[0.8, 0.45, np.nan, 0.1]})
#Just incase your NaN are strings as indicated by your original posts.
df=df.replace('NaN', np.nan)
df= df.set_index('location')
df1 = df.where(df<.33).reset_index()
df2 = df.where((df>.34) & (df<.65)).reset_index()
df3 = df.where(df>.66).reset_index()
print(df1)
print('\n')
print(df2)
print('\n')
print(df3)
輸出：
location Person_1 Person_2
0 4991 NaN NaN
1 8870 0.2 NaN
2 2703 NaN NaN
3 9674 NaN 0.1
location Person_1 Person_2
0 4991 NaN NaN
1 8870 NaN 0.45
2 2703 0.5 NaN
3 9674 NaN NaN
location Person_1 Person_2
0 4991 NaN 0.8
1 8870 NaN NaN
2 2703 NaN NaN
3 9674 0.7 NaN
添加回答
舉報(bào)
0/150
提交
取消