Python | Pandas Series.argsort()

Pandas series is a One-dimensional ndarray with axis labels. The labels need not be unique but must be a hashable type. The object supports both integer- and label-based indexing and provides a host of methods for performing operations involving the index.

Pandas Series.argsort() function returns the indices that would sort the underlying data of the given series object.

Syntax: Series.argsort(axis=0, kind=’quicksort’, order=None)

Parameter :
axis : Has no effect but is accepted for compatibility with numpy.
kind : {‘mergesort’, ‘quicksort’, ‘heapsort’}, default ‘quicksort’
order : Has no effect but is accepted for compatibility with numpy.

Returns : argsorted : Series, with -1 indicated where nan values are present

Example #1: Use Series.argsort() function to return the sequence of index which will sort the underlying data of the given series object.

# importing pandas as pd
import pandas as pd

# Creating the Series
sr = pd.Series([34, 5, 13, 32, 4, 15])

# Create the Index
index_ = ['Coca Cola', 'Sprite', 'Coke', 'Fanta', 'Dew', 'ThumbsUp']

# set the index
sr.index = index_

# Print the series
print(sr)

Output :

Coca Cola    34
Sprite        5
Coke         13
Fanta        32
Dew           4
ThumbsUp     15
dtype: int64

Now we will use Series.argsort() function to return a sequence of indices which will sort the underlying data of the given series object.

# return the indices which will
# sort the series
result = sr.argsort()

# Print the result
print(result)

# Let's sort the series using the result
print(sr[result])

Output :

Coca Cola    4
Sprite       1
Coke         2
Fanta        5
Dew          3
ThumbsUp     0
dtype: int64

Dew           4
Sprite        5
Coke         13
ThumbsUp     15
Fanta        32
Coca Cola    34
dtype: int64

As we can see in the output, the Series.argsort() function has successfully returned a series object containing the indices which will sort the given series object.

Example #2 : Use Series.argsort() function to return the sequence of index which will sort the underlying data of the given series object.

# importing pandas as pd
import pandas as pd

# Creating the Series
sr = pd.Series([11, 21, 8, 18, 65, 18, 32, 10, 5, 32, None])

# Create the Index
# apply yearly frequency
index_ = pd.date_range('2010-10-09 08:45', periods = 11, freq ='Y')

# set the index
sr.index = index_

# Print the series
print(sr)

Output :

2010-12-31 08:45:00    11.0
2011-12-31 08:45:00    21.0
2012-12-31 08:45:00     8.0
2013-12-31 08:45:00    18.0
2014-12-31 08:45:00    65.0
2015-12-31 08:45:00    18.0
2016-12-31 08:45:00    32.0
2017-12-31 08:45:00    10.0
2018-12-31 08:45:00     5.0
2019-12-31 08:45:00    32.0
2020-12-31 08:45:00     NaN
Freq: A-DEC, dtype: float64

Now we will use Series.argsort() function to return a sequence of indices which will sort the underlying data of the given series object.

# return the indices which will
# sort the series
result = sr.argsort()

# Print the result
print(result)

# Let's sort the series using the result
print(sr[result])

Output :

2010-12-31 08:45:00    8
2011-12-31 08:45:00    2
2012-12-31 08:45:00    7
2013-12-31 08:45:00    0
2014-12-31 08:45:00    3
2015-12-31 08:45:00    5
2016-12-31 08:45:00    1
2017-12-31 08:45:00    6
2018-12-31 08:45:00    9
2019-12-31 08:45:00    4
2020-12-31 08:45:00   -1
Freq: A-DEC, dtype: int64

2018-12-31 08:45:00     5.0
2012-12-31 08:45:00     8.0
2017-12-31 08:45:00    10.0
2010-12-31 08:45:00    11.0
2013-12-31 08:45:00    18.0
2015-12-31 08:45:00    18.0
2011-12-31 08:45:00    21.0
2016-12-31 08:45:00    32.0
2019-12-31 08:45:00    32.0
2014-12-31 08:45:00    65.0
2020-12-31 08:45:00     NaN
dtype: float64

As we can see in the output, the Series.argsort() function has successfully returned a series object containing the indices which will sort the given series object. Notice the function has returned -1 as the index position for the missing values.