Extract values in Pandas value_counts()

Question

Say we have used pandas dataframe[column].value_counts() which outputs:

 apple   5 
 sausage 2
 banana  2
 cheese  1

How do you extract the values in the order same as shown above from max to min ?

e.g: [apple,sausage,banana,cheese]

Could you please provide a [MCVE](http://stackoverflow.com/help/mcve)? Did you use `dataframe.value_counts()` or `series.value_counts()`? What datatype do you have the output in? — albert, Feb 20 '16 at 13:05
Note that the output of `value_counts()` is a series, so any series methods can be used, but often you'd just save it as is, depending on what you want to do with it later. — JohnE, Feb 20 '16 at 15:51

score 133 · Accepted Answer · answered Feb 20 '16 at 13:19

133

Try this:

dataframe[column].value_counts().index.tolist()
['apple', 'sausage', 'banana', 'cheese']

answered Feb 20 '16 at 13:19

Mike Müller

82,630
20
166
161

Awesome! On another note, when you print `dataframe[column].value_counts()` you get a dtype at the bottom. Is there a way not to get that? – JamesButterlips Feb 20 '16 at 13:22
28

Try: `dataframe[column].value_counts().to_frame()` – Mike Müller Feb 20 '16 at 13:36
6

one more option `.value_counts().index` and `.value_counts().values` – Aray Karjauv Jul 08 '20 at 18:24

score 56 · Answer 2 · answered Jan 05 '18 at 14:16

#!/usr/bin/env python

import pandas as pd

# Make example dataframe
df = pd.DataFrame([(1, 'Germany'),
                   (2, 'France'),
                   (3, 'Indonesia'),
                   (4, 'France'),
                   (5, 'France'),
                   (6, 'Germany'),
                   (7, 'UK'),
                   ],
                  columns=['groupid', 'country'],
                  index=['a', 'b', 'c', 'd', 'e', 'f', 'g'])

# What you're looking for
values = df['country'].value_counts().keys().tolist()
counts = df['country'].value_counts().tolist()

Now, print(df['country'].value_counts()) gives:

France       3
Germany      2
UK           1
Indonesia    1

and print(values) gives:

['France', 'Germany', 'UK', 'Indonesia']

and print(counts) gives:

[3, 2, 1, 1]

score 28 · Answer 3 · answered Apr 04 '18 at 12:46

28

If anyone missed it out in the comments, try this:

dataframe[column].value_counts().to_frame()

answered Apr 04 '18 at 12:46

Sawant

4,321
1
27
30

score 10 · Answer 4 · answered Jun 28 '20 at 14:08

10

The best way to extract the values is to just do the following

json.loads(dataframe[column].value_counts().to_json())

This returns a dictionary which you can use like any other dict. Using values or keys.

 {"apple": 5, "sausage": 2, "banana": 2, "cheese": 1}

answered Jun 28 '20 at 14:08

SummmerFort

369
3
8

`sorted(dict, key=lambda key: -dict[key])` – Androbin Jun 28 '20 at 14:46

score 2 · Answer 5 · answered Feb 20 '16 at 13:25

First you have to sort the dataframe by the count column max to min if it's not sorted that way already. In your post, it is in the right order already but I will sort it anyways:

dataframe.sort_index(by='count', ascending=[False])
    col     count
0   apple   5
1   sausage 2
2   banana  2
3   cheese  1

Then you can output the col column to a list:

dataframe['col'].tolist()
['apple', 'sausage', 'banana', 'cheese']

Extract values in Pandas value_counts()

5 Answers5

Linked

Related