ID rows containing values greater than corresponding values based on a criteria from another row - pandas

Question

I have a grouped dataframe. I have created a flag that identifies if values in a row are less than the group maximums. This works fine. However I want to unflag rows where the value contained in a third column is greater than the value in the same (third) column within each group. I have a feeling there shoule be an elegant and pythonic way to do this but I can't figure it out.
The flag I have shown in the code compares the maximum value of tour_duration within each hh_id to the corresponding value of "comp_expr" and if found less, assigns "1" to the column flag. However, I want values in the flag column to be 0 if min(arrivaltime) for each subgroup tour_id > max(arrivaltime) for the tour_id whose tour_duration is found to be maximum within each hh_id. For example, in the given data, tour_id 16300 has the highest value of tour_duration. But tour_id 16200 has min arrivaltime 1080 which is < max(arrivaltime) for tour_id 16300 (960). So flag for all tour_id 16200 should be 0.
Kindly assist.
import pandas as pd
import numpy as np
stops_data = pd.DataFrame({'hh_id': [20044,20044,20044,20044,20044,20044,20044,20044,20044,20044,20044,20122,20122,20122,20122,20122,20122,20122,20122,20122,20122,20122,20122,20122,],'tour_id':[16300,16300,16100,16100,16100,16100,16200,16200,16200,16000,16000,38100,38100,37900,37900,37900,38000,38000,38000,38000,38000,38000,37800,37800],'arrivaltime':[360,960,900,900,900,960,1080,1140,1140,420,840,300,960,780,720,960,1080,1080,1080,1080,1140,1140,480,900],'tour_duration':[600,600,60,60,60,60,60,60,60,420,420,660,660,240,240,240,60,60,60,60,60,60,420,420],'comp_expr':[1350,1350,268,268,268,268,406,406,406,974,974,1568,1568,606,606,606,298,298,298,298,298,298,840,840]})
stops_data['flag'] = np.where(stops_data.groupby(['hh_id'])
['tour_duration'].transform(max) < stops_data['comp_expr'],0,1)
This is my current output:Current dataset and output
This is my desired output, please see flag column: Desired output, see changed flag values in bold

Answer 1

>>> stops_data.loc[stops_data.tour_id
.isin(stops_data.loc[stops_data.loc[stops_data
.groupby(['hh_id','tour_id'])['arrivaltime'].idxmin()]
.groupby('hh_id')['arrivaltime'].idxmax()]['tour_id']), 'flag'] = 0
>>> stops_data
hh_id tour_id arrivaltime tour_duration comp_expr flag
0 20044 16300 360 600 1350 0
1 20044 16300 960 600 1350 0
2 20044 16100 900 60 268 1
3 20044 16100 900 60 268 1
4 20044 16100 900 60 268 1
5 20044 16100 960 60 268 1
6 20044 16200 1080 60 406 0
7 20044 16200 1140 60 406 0
8 20044 16200 1140 60 406 0
9 20044 16000 420 420 974 0
10 20044 16000 840 420 974 0
11 20122 38100 300 660 1568 0
12 20122 38100 960 660 1568 0
13 20122 37900 780 240 606 1
14 20122 37900 720 240 606 1
15 20122 37900 960 240 606 1
16 20122 38000 1080 60 298 0
17 20122 38000 1080 60 298 0
18 20122 38000 1080 60 298 0
19 20122 38000 1080 60 298 0
20 20122 38000 1140 60 298 0
21 20122 38000 1140 60 298 0
22 20122 37800 480 420 840 0
23 20122 37800 900 420 840 0