How to combine only specific rows in a dataframe

Question

There is 1 Dataframe and 1 series

A dataframe (df1) consists of a model, its parts and quantities.

df1 = pd.DataFrame({'model':['A','A','A','B','B','C','C','C','D','D'],
                'part':['ax','ay','az','bx','by','cx','cy','cz','dx','dy'],
                'qty':[1,2,3,2,4,2,5,3,1,2]})

Series(s1) has information about whether parts for each model can be merged or not. "on" if merging is possible, "off" if not possible

s1 = pd.Series({'A':'on', 'B':'off','C':'off','D':'on'})

Leave the model with merge off, A model with merge on wants to combine the rows and display the sum in qty.

The picture below is what I want

Answer 1

Here is one very direct approach:

1. join your "merge" indicator to your main dataframe
2. subset the part where you want to groupby on the model
3. do the groupby on the specified subset
4. join back to the rest of the data that you did not want to merge

Code below:

import pandas as pd
df1 = pd.DataFrame({'model':['A','A','A','B','B','C','C','C','D','D'],
            'part':['ax','ay','az','bx','by','cx','cy','cz','dx','dy'],
            'qty':[1,2,3,2,4,2,5,3,1,2]})
s1 = pd.Series({'A':'on', 'B':'off','C':'off','D':'on'}, name='merge?')

Part 1

df2 = df1.merge(s1,left_on=['model'], right_index=True)

Part 2

to_merge_ind = df2.loc[:,'merge?'] == 'on'

Part 3

merged = df1[to_merge_ind].groupby('model')\
         .agg({'part': lambda x: '/'.join(x), 'qty': sum}).reset_index()

Part 4

final = merged.append(df1[~to_merge_ind])

Output: