如何在 pandas 中映射列值? [英] How to map a column value in pandas?

问题描述

在我的文件中，一列包含不同的成绩(列名='成绩).

In my files one column containing different grades( columns name='Grades).

例如:91 50K，92 60K，DIESEL，ADBlU等.

eg: 91 50K,92 60K,DIESEL,ADBlU etc..

对于所有这些年级，我需要将其分类为几个年级；

For all these grades I need to categorize them in few grades;

例如: 91 50K= Petrol

在我的python中，我该怎么做?请注意，我可以将整个列传递给函数.但是函数必须用正确的值重写每一行的值；

In my python, how can i do this? Note that i can pass the whole column to the function. but function has to rewrite each row's value with the correct one;

def checkgrades(data):
df['Grades']=???
???

按照下面的答案，我尝试；

As per below answers I try;

 df_dips=map_dips_grades(df_dips)
df_sales = df_sales.merge(df_dips, left_on=['Site Name', 'Date','GradeNo'],
                              right_on=['Site', 'Dip Time', 'Product'], how='left').fillna(0)



def map_dips_grades(data):

    d1 = {'Diesel': ['DIESEL', 'DIESEL 1'],
          'Unleaded': ['91','91 UNLEADED'],
          'PULP':['95','95 ULP'],
          'PULP98':['98','98 20K'],
          'Vortex Diesel':['DIESEL ULT R'],
          'Adblue':['ADBLU','ADO']}

    d = {k: oldk for oldk, oldv in d1.items() for k in oldv}

    data['Product'].map(d)
    return data

但是我明白了

ValueError: You are trying to merge on int64 and object columns. If you wish to proceed you should use pd.concat

推荐答案

您可以在Grades中创建所有可能值的字典，然后在

You can create dictionary of all possible values in Grades and then Series.map:

#test all possible unique values
print (df['Grades'].unique())

d = {'91 50K':'Petrol','92 60K':'Petrol','DIESEL':'Diesel',...}

df['Grades'] = df['Grades'].map(d)

另一种可能减少打字的字典是列表字典:

Another possible dictionary for less typing is dict of lists:

d1 = {'Petrol':['91 50K','92 60K'],
      'Diesel':['DIESEL']}

#swap key values in dict
#http://stackoverflow.com/a/31674731/2901002
d = {k: oldk for oldk, oldv in d1.items() for k in oldv}
print (d)
{'91 50K': 'Petrol', '92 60K': 'Petrol', 'DIESEL': 'Diesel'}

df['Grades'] = df['Grades'].map(d)

这篇关于如何在 pandas 中映射列值?的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！