我们如何计算数据框中的项目并将结果分配给数据框中的新列？

Question

I have street addresses that look like this.

250 EAST HOUSTON STREET
211 EAST 3RD STREET
182 EAST 2ND STREET
511 EAST 11TH STREET
324 EAST 4TH STREET
324 EAST 4TH STREET
324 EAST 4TH STREET
324 EAST 4TH STREET
324 EAST 4TH STREET
324 EAST 4TH STREET
324 EAST 4TH STREET
324 EAST 4TH STREET
754 EAST 6TH STREET

How can I get counts, in the same data frame, like this?

250 EAST HOUSTON STREET 3
211 EAST 3RD STREET     1
182 EAST 2ND STREET     1
511 EAST 11TH STREET    1
324 EAST 4TH STREET     8
324 EAST 4TH STREET     8
324 EAST 4TH STREET     8
324 EAST 4TH STREET     8
324 EAST 4TH STREET     8
324 EAST 4TH STREET     8
324 EAST 4TH STREET     8
324 EAST 4TH STREET     8
754 EAST 6TH STREET     1

The name of the field that I want to count is 'Street'. I found some code that counts dupes, but it does a group by and takes everything into a new data frame. I want to assign counts to a new column in the same data frame. Thanks!

Answer 1

Use groupby + transform . Transform allows you to call the series to a new column. If you do not use transform, then you have a consolidated series that is a mismatch for the dataframe and your column will instead be filled with NaN values:

import pandas as pd
# df = pd.read_clipboard('\s\s+', header=None).rename({0: 'Street'}, axis=1) # how I read in your data from your StackOverflow question
df['Count'] = df.groupby('Street')['Street'].transform('count')
df
Out[1]: 
                     Street  Count
0   250 EAST HOUSTON STREET  1
1       211 EAST 3RD STREET  1
2       182 EAST 2ND STREET  1
3      511 EAST 11TH STREET  1
4       324 EAST 4TH STREET  8
5       324 EAST 4TH STREET  8
6       324 EAST 4TH STREET  8
7       324 EAST 4TH STREET  8
8       324 EAST 4TH STREET  8
9       324 EAST 4TH STREET  8
10      324 EAST 4TH STREET  8
11      324 EAST 4TH STREET  8
12      754 EAST 6TH STREET  1

Answer 2

I'm really new to trying to help people on this platform so feel free to tell me I'm not looking in the right places for the info, but are you using Pandas or another library? If you're using Pandas I think there's a method called valuecount (maybe value_count) that could be useful. Sorry I can't be more helpful but I'm learning the ropes here.