The Pandas Dataframe.Where () function explained

The Python Pandas function DataFrame.where() is used to perform Conditional data handling In data. It allows programmers to replace or hide values ​​in a dataframa pandas based on a specific condition.

Pandas syntax DataFrame.where()

The function where() accepted up to five settings and follows the basic syntax shown below:

DataFrame.where(cond, other=nan, inplace=False, axis=None, level=None)

In this case, the function is applied to a dataframe and only the values ​​which fulfill the specified condition (cond) remain unchanged. All other values ​​are replaced by the values ​​specified in other.

Relevant parameters

Pandas DataFrame.where() Accepts different parameters that allow flexible adaptation of data handling:

Pandas application DataFrame.where()

The function where() can be used in many situations where Conditional data handling are necessary. For example, this is data cleaning or Creation of new columns based on conditions.

Conditional replacement of values

Suppose you have a dataframa containing the sales results of a company and that you want to display only positive results. All negative results must be replaced by 0. This can be done with Pandas DataFrame.where(). First of all, a dataaframa is created:

import pandas as pd
# Création d’un DataFrame d’exemple
data = {
    'Région': ['Nord', 'Sud', 'Est', 'Ouest'],
    'Ventes_Q1': [15000, -5000, 3000, -1000],
    'Ventes_Q2': [20000, 25000, -7000, 5000]
}
df = pd.DataFrame(data)
print(df)

The above code returns the following dataframa:

Région    Ventes_Q1    Ventes_Q2
0    Nord         15000         20000
1    Sud          -5000         25000
2    Est            3000         -7000
3    Ouest        -1000          5000

By calling where()you can now replace all negative values ​​with 0. To do this, you must make sure that only columns containing digital values ​​are taken into accountotherwise the comparison will not work.

# Remplacement conditionnel de valeurs
df_positive = df.copy()
df_positive[['Ventes_Q1', 'Ventes_Q2']] = df[['Ventes_Q1', 'Ventes_Q2']].where(df[['Ventes_Q1', 'Ventes_Q2']] > 0, 0)
print(df_positive)

The resulting dataframa df_positive only contains positive sales results and replaces all negative values ​​with 0 As desired:

Région    Ventes_Q1    Ventes_Q2
0    Nord         15000         20000
1    Sud              0          25000
2    Est          3000                0
3    Ouest          0            5000

Conditional masking of values

Pandas DataFrame.where() Can also be used to hide values, that is to say only make visible certain parts of a dataframa. In what follows, the dataframe should only display the values ​​higher than a certain threshold (in this case, 10000). Here again, you should make sure that only digital columns are taken into account:

# N’afficher que les valeurs supérieures à 10 000
df_masked = df.copy()
df_masked[['Ventes_Q1', 'Ventes_Q2']] = df[['Ventes_Q1', 'Ventes_Q2']].where(df[['Ventes_Q1', 'Ventes_Q2']] > 10000)
print(df_masked)

In this case, the resulting dataframa df_masked only displayed the values ​​higher than 10000. All other values ​​are displayed as NaN ::

Région    Ventes_Q1    Ventes_Q2
0    Nord         15000.0     20000.0
1    Sud              NaN      25000.0
2    Est              NaN            NaN
3    Ouest          NaN            NaN

The function where() De Pandas is therefore a powerful tool for filtering, transforming and cleaning data effectively in a dataaframa, while preserving its original structure.