.reset_index()
Through the course of exploratory analysis, and other data work, a DataFrame object will often be modified to clean and/or restructure the data. Through this work an index may become discontinuous or additional levels may be added or subtracted from the index. The .reset_index() method can be used to reestablish a continuous index as well as remove one or more unwanted levels.
Syntax
df = dataframe_value.reset_index()The .reset_index() method provides the following parameters:
- level: Takes integer, string, tuple, list, or None values and is set to None by default. Removes the given levels from the index, by default all levels are removed.
- drop: Takes a boolean value and is set to
Falseby default. When this parameter is set toTrueit replaces the previous DataFrame index with the new index provided by.reset_index(), otherwise it sets the new index in front of the old index. - inplace: Takes a boolean value and is set to
Falseby default. When this parameter is set toTrueit applies all changes to the current instance of the DataFrame, otherwise it creates a new DataFrame instance with the changes applied to that DataFrame. - col_level: Takes integer or string values and is set to 0 by default. Determines what level the labels are inserted into when the columns have multiple levels. The first level is set by default.
- col_fill: Takes a string, list, or None and is set to None by default. Determines how the other levels are named when the columns have multiple levels. Uses the index name by default.
- allow_duplicates: Optional parameter which takes a boolean value and is set to
lib.no_defaultby default. When this parameter is set toTrueit allows duplicate column labels to be created. - names: Takes integer, string, 1-dimensional list, or None values and is set to None by default. Renames the index DataFrame column. In the case that the Dataframe has a MultiIndex this value has to be a list or tuple equal in length to the number of levels.
Actions Which Cause Indexing Issues
Common examples include but are not limited to:
- Changing the order of columns.
- Using a column or multiple columns of the database as an index instead of the default index pandas provides and then later needing a numbered index.
- Filtering out certain rows of the DataFrame based on the value of a column of data.
- Changing the order of rows of data by sorting those rows by the values of a certain column of data such as Date or Employee ID Number.
- Adding columns and or rows of data to the DataFrame that were not present in the original DataFrame.
Example
To follow along a copy of the Austin_Animal_Center_intakes.csv can be downloaded from city of Austin data portal.
The Original DataFrame
import pandas as pddf = pd.read_csv('Austin_Animal_Center_intakes.csv').head()pd.set_option('display.max_columns', None)print(df)
This results in the following output:
Animal ID Name DateTime MonthYear \0 A786884 *Brock 01/03/2019 04:19:00 PM January 20191 A706918 Belle 07/05/2015 12:59:00 PM July 20152 A724273 Runster 04/14/2016 06:43:00 PM April 20163 A857105 Johnny Ringo 05/12/2022 12:23:00 AM May 20224 A682524 Rio 06/29/2014 10:38:00 AM June 2014Found Location Intake Type Intake Condition \0 2501 Magin Meadow Dr in Austin (TX) Stray Normal1 9409 Bluegrass Dr in Austin (TX) Stray Normal2 2818 Palomino Trail in Austin (TX) Stray Normal3 4404 Sarasota Drive in Austin (TX) Public Assist Normal4 800 Grove Blvd in Austin (TX) Stray NormalAnimal Type Sex upon Intake Age upon Intake \0 Dog Neutered Male 2 years1 Dog Spayed Female 8 years2 Dog Intact Male 11 months3 Cat Neutered Male 2 years4 Dog Neutered Male 4 yearsBreed Color0 Beagle Mix Tricolor1 English Springer Spaniel White/Liver2 Basenji Mix Sable/White3 Domestic Shorthair Orange Tabby4 Doberman Pinsch/Australian Cattle Dog Tan/Gray
Removing Cats From the Dataframe
# This section of code removes the furball from our dog DataFramedf = df[df['Animal Type'] != 'Cat']# Uncommenting the line below this line will remove the index of the original DataFrame and reset the order# df.reset_index(inplace = True, drop = True)print(df)
This is the output without df.reset_index(inplace = True, drop = True):
Animal ID Name DateTime MonthYear \0 A786884 *Brock 01/03/2019 04:19:00 PM January 20191 A706918 Belle 07/05/2015 12:59:00 PM July 20152 A724273 Runster 04/14/2016 06:43:00 PM April 20164 A682524 Rio 06/29/2014 10:38:00 AM June 2014Found Location Intake Type Intake Condition \0 2501 Magin Meadow Dr in Austin (TX) Stray Normal1 9409 Bluegrass Dr in Austin (TX) Stray Normal2 2818 Palomino Trail in Austin (TX) Stray Normal4 800 Grove Blvd in Austin (TX) Stray NormalAnimal Type Sex upon Intake Age upon Intake \0 Dog Neutered Male 2 years1 Dog Spayed Female 8 years2 Dog Intact Male 11 months4 Dog Neutered Male 4 yearsBreed Color0 Beagle Mix Tricolor1 English Springer Spaniel White/Liver2 Basenji Mix Sable/White4 Doberman Pinsch/Australian Cattle Dog Tan/Gray
The indexing now jumps from two to four after the row containing the cat is removed. This can become very messy when dealing with large DataFrames containing hundreds or even thousands of rows.
This is the output with df.reset_index(inplace = True, drop = True):
Animal ID Name DateTime MonthYear \
0 A786884 *Brock 01/03/2019 04:19:00 PM January 2019
1 A706918 Belle 07/05/2015 12:59:00 PM July 2015
2 A724273 Runster 04/14/2016 06:43:00 PM April 2016
3 A682524 Rio 06/29/2014 10:38:00 AM June 2014
Found Location Intake Type Intake Condition \
0 2501 Magin Meadow Dr in Austin (TX) Stray Normal
1 9409 Bluegrass Dr in Austin (TX) Stray Normal
2 2818 Palomino Trail in Austin (TX) Stray Normal
3 800 Grove Blvd in Austin (TX) Stray Normal
Animal Type Sex upon Intake Age upon Intake \
0 Dog Neutered Male 2 years
1 Dog Spayed Female 8 years
2 Dog Intact Male 11 months
3 Dog Neutered Male 4 years
Breed Color
0 Beagle Mix Tricolor
1 English Springer Spaniel White/Liver
2 Basenji Mix Sable/White
3 Doberman Pinsch/Australian Cattle Dog Tan/GrayAfter applying df.reset_index(inplace = True, drop = True) the DataFrame index order is now neat and continuous for easy indexing.
