Python Pandas: How can I group by and assign an id to all the items in a group?

问题

I have df:

domain           orgid
csyunshu.com    108299
dshu.com        108299
bbbdshu.com     108299
cwakwakmrg.com  121303
ckonkatsunet.com    121303

I would like to add a new column with replaces domain column with numeric ids per orgid:

domain           orgid   domainid
csyunshu.com    108299      1
dshu.com        108299      2
bbbdshu.com     108299      3
cwakwakmrg.com  121303      1
ckonkatsunet.com 121303     2

I have already tried this line but it does not give the result I want:

df.groupby('orgid').count['domain'].reset_index()

Can anybody help?

回答1:

You can call rank on the groupby object and pass param method='first':

In [61]:
df['domainId'] = df.groupby('orgid')['orgid'].rank(method='first')
df

Out[61]:
             domain   orgid  domainId
0      csyunshu.com  108299         1
1          dshu.com  108299         2
2       bbbdshu.com  108299         3
3    cwakwakmrg.com  121303         1
4  ckonkatsunet.com  121303         2

If you want to overwrite the column you can do:

df['domain'] = df.groupby('orgid')['orgid'].rank(method='first')

回答2:

You can use LabelEncoder from sklearn.preprocessing like :

df["domain"] = LabelEncoder().fit_transform(df.domain)

来源：https://stackoverflow.com/questions/36063251/python-pandas-how-can-i-group-by-and-assign-an-id-to-all-the-items-in-a-group