Search values in dataframe by condition on another column

Question

I need to get the value in column A for the closest value in column B for each multiple of ‘trigger’

for instance, in the dataframe below :

import random

trigger = 100

info2 = {'A': [0]*100,'B': [0]*100}
dfA = pd.DataFrame(info2)


for i in range(1, len(dfA)):
    dfA.loc[i,'B'] = i*3.78
    dfA.loc[i,'A'] = i*10

    
dfA

Since the closest value to trigger1 would be 98.28 from row n°26
The closest value to trigger2 would be 200.34 from row n°53
The closest value to trigger*3 would be 298.62 from row n°79

The expected result would be :
result = [260,530,790]

Asked By: JK2018

||

Source

Answer 1

This could do

import numpy as np
triggers = {'100': 100, '200': 200, '300': 300}
for k, v in triggers.items():
    dfA['delta_val'] = np.abs(dfA['B'] - v)
    triggers[k] = dfA[dfA.delta_val == dfA.delta_val.min()]['A'].values[0]
print(triggers)

# {'100': 260, '200': 530, '300': 790}

Answered By: TommyLeeJones

Answer 2

Another approach is :

import pandas as pd
import numpy as np

trigger = 100

info2 = {'A': [0]*100,'B': [0]*100}
dfA = pd.DataFrame(info2)

for i in range(1, len(dfA)):
    dfA.loc[i,'B'] = i*3.78
    dfA.loc[i,'A'] = i*10

result = []
for t in np.arange(trigger, trigger*4, trigger):
    idx = (np.abs(dfA['B'] - t)).idxmin()
    result.append(dfA.loc[idx, 'A'])

print(result)

which gives what you expected.

Answered By: Serge de Gosson de Varennes

Answer 3

difference = abs(dfA - target)
min_index = difference.sum(axis=1).idxmin()
result = dfA.loc[min_index, :]
print(result)

Answered By: keyur

Answer 4

Use merge_asof:

pd.merge_asof(pd.Series(np.arange(trigger, dfA['B'].max(), trigger), name='B'),
              dfA, on='B', direction='nearest')

NB. dfA must first be sorted on B.

Output:

       B    A
0  100.0  260
1  200.0  530
2  300.0  790

If you also want the value of B:

pd.merge_asof(pd.Series(np.arange(trigger, dfA['B'].max(), trigger), name='trigger'),
              dfA, left_on='trigger', right_on='B', direction='nearest')

Output:

   trigger    A       B
0    100.0  260   98.28
1    200.0  530  200.34
2    300.0  790  298.62

Answered By: mozway

Search values in dataframe by condition on another column

Question:

Answers: