Python | Pandas Series.argsort()

With the help of Pandas Series.argsort(), one can sort the elements of series in pandas. But the main thing in pandas series is we get the output as index values of the sorted elements in series. In the later code demonstration, we will explain that how we get the output as sorted index values.

Syntax: pandas.Series.argsort(axis=0, kind=’quicksort’, order=None) Parameters: axis : It is useful for numpy. kind : {‘mergesort’, ‘quicksort’, ‘heapsort’}, default ‘quicksort’ order : It is useful for numpy. Returns: argsorted Series, with -1 indicated where nan values are present

To get the link to csv file, click on nba.csv Code #1 : In this code you will see that we are taking a simple series of some integer values and try to sort on the basis of different methods of sorting algorithms like quicksort, mergesort and heapsort but by default it will assume as quicksort. Lets see the code below and the following output. 

Python3

# importing pandas
import pandas as pd

# reading the csv
data = pd.read_csv("nba.csv")

data.dropna(inplace = True)

# creating series form weight column
g = pd.Series(data['Weight'].head())
print(g)

gfg = g.argsort(axis = 0, kind ='quicksort', order = None)

print(gfg)

Output:

0    180.0
1    235.0
3    185.0
6    235.0
7    238.0
Name: Weight, dtype: float64
0    0
1    2
3    1
6    3
7    4
Name: Weight, dtype: int64

As you can see in the output and it looks strange that instead of getting the sorted values in series why we got these numbers. This is the major concept of Series.argsort() method it return the index value of a smallest number first and index value of largest value in the end. As we have 1 is the smallest number and it’s index value is 4 then 4 will come first and this concept as flow as following output. Code #2 :

Python3

# importing pandas
import pandas as pd

# reading the csv
data = pd.read_csv("nba.csv")

data.dropna(inplace = True)

# creating series form weight column
g = pd.Series(data['Weight'].head())
print(g)

gfg = g.argsort(axis = 0, kind ='mergesort', order = None)

print(gfg)

Output:

0    180.0
1    235.0
3    185.0
6    235.0
7    238.0
Name: Weight, dtype: float64
0    0
1    2
3    1
6    3
7    4
Name: Weight, dtype: int64

Code #3 :

Python3

# importing pandas
import pandas as pd

# reading the csv
data = pd.read_csv("nba.csv")

data.dropna(inplace = True)

# creating series form weight column
g = pd.Series(data['Weight'].head())
print(g)

gfg = g.argsort(axis = 0, kind ='heapsort', order = None)

print(gfg)

Output:

0    180.0
1    235.0
3    185.0
6    235.0
7    238.0
Name: Weight, dtype: float64
0    0
1    2
3    1
6    3
7    4
Name: Weight, dtype: int64

What is the output when we have missing values ? As we have explained above that if we want to handle the missing values then in the place of None it will give the output as -1. 

Python3

import pandas as pd

# importing pandas
import pandas as pd

# reading the csv
data = pd.read_csv("nba.csv")

# creating series form weight column
g = pd.Series(data['Weight'])
print(g)

gfg = g.argsort(axis = 0, kind ='mergesort', order = None)

print(gfg)

Output:

450    226.0
451    206.0
452    234.0
453    203.0
454    179.0
455    256.0
456    231.0
457      NaN
Name: Weight, Length: 458, dtype: float64
450    237
451     41
452    188
453    395
454    330
455    302
456    405
457     -1
Name: Weight, Length: 458, dtype: int64