Python CSV 分析的数据

Question

我将一个 csv 转换为一个名为 a 的列表。我有一种方法可以通过条件对我的数据进行分类。问题是它不起作用。如果在我的所有Cliente 上都有任何称为“稳定”的元素，我将条件设置为'Estable'，这不是我需要的，但对于所有没有'Estable' 的客户作为 AAA 和 BBB，我希望你把'NoAnalyzed' 我在代码下面解释。

import pandas as pd

a = [['Cliente', 'Fecha', 'Variables', 'Dia Previo', 'Mayor/Menor', 'Dia a Analizar', 'Analisis'], 
['AAA', '27/12/2017', 'ECPM_medio', '0.41', 'Dentro del Margen', '0.35', 'Incremento'],
['BBB', '27/12/2017', 'ECPM_medio', '1.06', 'Dentro del Margen', '1.06', 'Alerta'],
['CCC', '27/12/2017', 'ECPM_medio', '1.06', 'Dentro del Margen', '1.06', 'Estable']]



headers = a.pop(0)
df = pd.DataFrame(a, columns = headers)
df['Analisis']


for elemento in df['Analisis']:
    if elemento == 'Estable':
        df['Status'] = 'Stable: The client''s performance was Stable'
    else:
        df['Status'] = 'NoAnalyzed'


df1= df.groupby(['Cliente','Fecha', 'Status']).size()
df1

output:
>>>
Cliente  Fecha       Status                                    
AAA      27/12/2017  Stable: The clients performance was Stable    1
BBB      27/12/2017  Stable: The clients performance was Stable    1
CCC      27/12/2017  Stable: The clients performance was Stable    1

I need:
>>>
Cliente  Fecha       Status                                    
AAA      27/12/2017  NoAnalyzed    1
BBB      27/12/2017  NoAnalyzed    1
CCC      27/12/2017  Stable: The clients performance was Stable    1

Answer 1

我相信你需要numpy.where 或map，因为在pandas 中最好避免循环，因为速度慢：

mask =  df['Analisis'] == 'Estable'
df['Status'] = np.where(mask, 'Stable: The client''s performance was Stable', 'NoAnalyzed')

或类似的：

d = {True: 'Stable: The client''s performance was Stable',False: 'NoAnalyzed'}
df['Status'] = mask.map(d)

print (df)
  Cliente       Fecha   Variables Dia Previo        Mayor/Menor  \
0     AAA  27/12/2017  ECPM_medio       0.41  Dentro del Margen   
1     BBB  27/12/2017  ECPM_medio       1.06  Dentro del Margen   
2     CCC  27/12/2017  ECPM_medio       1.06  Dentro del Margen   

  Dia a Analizar    Analisis                                      Status  
0           0.35  Incremento                                  NoAnalyzed  
1           1.06      Alerta                                  NoAnalyzed  
2           1.06     Estable  Stable: The clients performance was Stable  

Answer 2

问题是您直接将单个值分配给列而不是列表/数组/系列。单个值在每一行中复制自身。我建议您制作一个列表并将其分配给您的 df['Status'] 列。

status=[]
for elemento in df['Analisis']:
    if elemento == 'Estable'
        status.append('Stable: The client''s performance was Stable')
    else:
        status.append('NoAnalyzed')

df['Status'] = status

这应该可行。