CODEWITHSUNDEEP
pythonpandasipywidgets
Nivi
Nivi

Reputation: 1137

Filtering pandas dataframe using ipywidgets

I am using a dropdown widget for the values in one of the columns in my dataframe. Based on the value I choose in the dropdown, I want the dataframe to filter only those rows. I would then go on to use the filtered dataframe for my analysis.

col1 |  col2    |  col3
  A  |   Orange | 2
  B  |   Apple  | 3
  A  |   Apple  | 4

I would have a dropdown (using ipywidgets) for col2 and when I choose Apple, I want to get the resulting dataframe.

col1  | col2   | col3
  B   |  Apple | 3
  A   |  Apple | 4

Upvotes: 0

Views: 5371

Answers (2)

Hestaron
Hestaron

Reputation: 302

This shows the fruit you select from the dropdown menu. I would advise you to watch the Scipy 2020 recoring on Interactive widgets if you want to learn more about widgets.

from ipywidgets import interact
import pandas as pd

def show_df(value):
    data = [['A', 'Orange', 2],['B', 'Apple', 3],['A', 'Apple', 4]]
    df = pd.DataFrame(data, columns=['Letter','Fruit','Number'])
    print(df[df.Fruit==value])

interact(show_df, value = ['Apple','Orange'])

There are more dedicated ways to create a dropdown menu but this works. Such as in this example:

import ipywidgets as widgets
widgets.Select(
    options=['Linux', 'Windows', 'OSX'],
    value='OSX',
    # rows=10,
    description='OS:',
    disabled=False
)

Upvotes: 2

Fantailed
Fantailed

Reputation: 118

I don't know about ipywidgets specifically but assuming you can get the value of your selection, say "Apple", filtering out by column values is quite easy.

df = pd.DataFrame({"col1": ["A", "B", "A"],
               "col2": ["Orange", "Apple", "Apple"],
               "col3": [2, 3, 4]})

df = df[df["col2"] == "Apple"]

Then, when you print df, you will get

  col1   col2  col3
1    B  Apple     3
2    A  Apple     4

Basically you create a mask (a dataframe of True and False values) through df["col2"] == "Apple" and then you can use that mask as an index to only get the values where the mask is True.

Upvotes: 0

Related Questions