df.mean是不是系列的真正意思？

我正在調試並遇到以下奇怪的行爲。我正在計算包含所有完全相同數字的熊貓系列的平均值。但是，pd.mean()給出了不同的數字。df.mean是不是系列的真正意思？

question1：爲什麼這個系列的意思是不同的數字？

問題2：tmm[-1]== tmm.mean()現在給出False。任何方式來忽略這個小的差異，並使結果爲真？我不喜歡abs(tmm[-1]-tmm.mean()) < xxx

方法，因爲不知道如何定義xxx。

import pandas as pd 
import decimal 

tmm = pd.Series(14.9199999999999999289457264239899814128875732421875, 
       index=range(30)) 
for t in tmm: 
    print(decimal.Decimal(t)) 
print('mean is') 
print(decimal.Decimal(tmm.mean()))

結果：

14.9199999999999999289457264239899814128875732421875 
14.9199999999999999289457264239899814128875732421875 
14.9199999999999999289457264239899814128875732421875 
14.9199999999999999289457264239899814128875732421875 
14.9199999999999999289457264239899814128875732421875 
14.9199999999999999289457264239899814128875732421875 
14.9199999999999999289457264239899814128875732421875 
14.9199999999999999289457264239899814128875732421875 
14.9199999999999999289457264239899814128875732421875 
14.9199999999999999289457264239899814128875732421875 
14.9199999999999999289457264239899814128875732421875 
14.9199999999999999289457264239899814128875732421875 
14.9199999999999999289457264239899814128875732421875 
14.9199999999999999289457264239899814128875732421875 
14.9199999999999999289457264239899814128875732421875 
14.9199999999999999289457264239899814128875732421875 
14.9199999999999999289457264239899814128875732421875 
14.9199999999999999289457264239899814128875732421875 
14.9199999999999999289457264239899814128875732421875 
14.9199999999999999289457264239899814128875732421875 
14.9199999999999999289457264239899814128875732421875 
14.9199999999999999289457264239899814128875732421875 
14.9199999999999999289457264239899814128875732421875 
14.9199999999999999289457264239899814128875732421875 
14.9199999999999999289457264239899814128875732421875 
14.9199999999999999289457264239899814128875732421875 
14.9199999999999999289457264239899814128875732421875 
14.9199999999999999289457264239899814128875732421875 
14.9199999999999999289457264239899814128875732421875 
14.9199999999999999289457264239899814128875732421875 
mean is 
14.9200000000000034816594052244909107685089111328125

來源

2016-08-14 Lisa

'float's本質上容易出現精度誤差，這就是爲什麼你會得到一個不同的意思。 –

「我不喜歡'abs（tmm [-1] -tmm.mean（）） BrenBarn

嘗試使用np.isclose()

tmm[20]== tmm.mean() 

False 

np.isclose(tmm[20], tmm.mean()) 
True

來源

2016-08-15 01:18:10 Merlin

回答你的兩個問題基本上是這樣的：

import pandas as pd 
import decimal 

tmm = pd.Series(decimal.Decimal(14.9199999999999999289457264239899814128875732421875), 
       index=range(30)) 
for t in tmm: 
    print(decimal.Decimal(t)) 
print('mean is') 
print(decimal.Decimal(tmm.mean()))

請確保您使用decimal.Decimal構造函數當你創建TMM，這是相當許多。

來源

2016-08-14 22:57:18 BPL

df.mean是不是系列的真正意思？

回答

相關問題