Скачать или смотреть Twitch Interview Question: Viewers Turned Streamers

Viewers Turned Streamers

Return the number of streamer sessions for each user whose very first session was as a viewer.

Include the user ID and count of streamer sessions for users whose earliest session (by session_start) was a 'viewer' session, regardless of whether they ever had a streamer session later. Sort the results by streamer session count in descending order, then by user ID in ascending order.

DataFrame: twitch_sessions

user_id: int64
session_start: datetime64[ns]
session_end: datetime64[ns]
session_id: int64
session_type: object

Code:
import pandas as pd
df=twitch_sessions
df1=df.loc[df.groupby('user_id')['session_start'].idxmin()]
df2=df1[(df1['session_type']=='viewer')]['user_id']
df1= df[(df['user_id'].isin(df2)) & (df['session_type']=='streamer')].groupby('user_id')['session_id'].count().to_frame('nsessions').reset_index()
df1.sort_values(by=['nsessions','user_id'],ascending=[False,True])
df1

Want more details, Kindly follow me on medium:   / aggarwalakshima  

#twitch #latest #level #hard #interview #question #python #stratascratch #dataframes #pandas #share #like #subscribe