DataFrame pandas: indexing – ionos

The indexing of data in Pandas Python allows effective and direct access to specific data within a dataframa. The use of a Pandas DataFrame index allows you to select specific lines and columns, which can considerably facilitate data analysis.

What's going on during indexing?

The indexing of a dataframa pandas aims to Facilitate the selection of specific dataframe elements. We can thus select lines and columns depending on their positions or their labels. Indexes can help find and process data faster by providing A kind of « address system » for the data structure.

Pandas syntax DataFrame.index

You can see the index values ​​of a dataframe pandas with the property index. The syntax is as follows:

Dataframas indexing syntax

There are several ways to index dataframes pandas. The indexing syntax varies depending on the desired operation.

Indexing with labels (column names)

DataFrames pandas can use column names for indexing. To do this, we will first create an example of DataFrame:

import pandas as pd
# Création d'un DataFrame d'exemple
données = {
    'Nom': ['Alice', 'Bob', 'Charlie'],
    'Âge': [25, 30, 35],
    'Ville': ['New York', 'Los Angeles', 'Chicago']
}
df = pd.DataFrame(données)
print(df)

Dataframa presents itself as follows:

Nom  Âge     Ville
0    Alice   25     New York
1      Bob   30  Los Angeles
2  Charlie   35      Chicago

If you want to access all the values ​​of a specific column, you can use its name in combination with the operator []. Just specify the name of the column in the indexing operator in the form of a chain (string) Python:

# Accès à la colonne « Âge »
print(df['Âge'])

You get the different age values ​​as a result:

0    25
1    30
2    35
Name: Âge, dtype: int64

If you are not interested in one, but in several columns, you can simply specify their names, separated by commas, in the indexing operator.

Indexing with loc[] (line name)

If you want to access a specific line of your dataframe, you need the Pandas Indexer loc[]. You then pass either the name of the line or the line number. In the following code example, we consider the same dataframe as before; We want to extract the front line containing the values ​​for « Alice »:

As expected, the values ​​corresponding to « Alice » are visible in the result:

Nom         Alice
Âge            25
Ville    New York
Name: 0, dtype: object

Indexing with iloc[] (lines of lines and columns)

Another way to access specific elements of your dataaframa is to use line and column numbers. To work with the digital index of the dataframe pandas, you need the DataFrame Iloc property[].

# Accès à la première ligne
print(df.iloc[0])
# Accès à la valeur dans la première ligne et la deuxième colonne
print(df.iloc[0, 1])

The results of using iloc[] look like this and refer the expected values:

Nom         Alice
Âge            25
Ville    New York
Name: 0, dtype: object

Index individual values

If you are only interested in a specific value of your dataframe, the indexor at is an effective way to extract this value. Simply define the line and column in which the value must be found, with their names. Thus, if the place of residence of Bob is interesting, we need the « city » column and the first line:

print(df.at(1, 'Ville'))

As requested, the exit is the city of residence of Bob, ie « Los Angeles ».

You can also use the Indexer iatwhich works in the same way as atbut who awaits the position instead of the name. The same result as in the example of previous code is obtained with the use of iat ::

print(df.iat[1, 2])

Boolean indexation

It is possible to index sub-assemblies of a dataframa based on a particular condition. In this case, we are talking about Boolean indexing. The condition to be verified must be assessed at True or at False and is placed directly in the indexing operator. To extract only the lines in which the person is over 30 years old, we can proceed as follows:

# Sélection des lignes où l'âge est supérieur à 30
print(df[df['Âge'] > 30])

The above condition only applies to « Charlie », 35 years old. The outing is therefore as follows:

Nom  Âge    Ville
2  Charlie   35  Chicago

Indexing is a fundamental pandas tool which allows effectively access to data and to extract sub-assemblies relevant for analysis.