How to save previous result that matched certain condition in pandas

user_5766952 0
Viewed 46

I am wondering how to save previous result, that matched certain conditions (df['condition'] in every row that follows. I know how to do it with for loop, but I know I should avoid them when working with pandas.

Below is an example. column df['desired_result] represents what I want to achieve.

import pandas as pd
import numpy as np

dates = pd.date_range('1/1/2000', periods=10)
values = np.arange(10.0, 20.0, 1.0)
data = {'date': dates, 'value': values}
df = pd.DataFrame.from_dict(data)

df['condition'] = [False, False, True, True, False, True, False, False, True, False]
df_valid = df[df['condition']]
df['desired_result'] = [np.nan, np.nan, 12, 13, 13, 15, 15, 15, 18, 18]
pandas vectorization dataframe python
1 Answers
# use df.where based on your condition and assign it to a new col
# Anywhere column condition is True return the value else return NaN
# then add ffill to forward fill NaN values

df['r'] = df['value'].where(df['condition'] == True, np.nan).ffill()

        date  value  condition  desired_result     r
0 2000-01-01   10.0      False             NaN   NaN
1 2000-01-02   11.0      False             NaN   NaN
2 2000-01-03   12.0       True            12.0  12.0
3 2000-01-04   13.0       True            13.0  13.0
4 2000-01-05   14.0      False            13.0  13.0
5 2000-01-06   15.0       True            15.0  15.0
6 2000-01-07   16.0      False            15.0  15.0
7 2000-01-08   17.0      False            15.0  15.0
8 2000-01-09   18.0       True            18.0  18.0
9 2000-01-10   19.0      False            18.0  18.0
user_91778770
Related