pandas-group-by.py

#!/usr/bin/env python3
#-*- coding:utf-8 -*-




import random
import os
import binascii



import pandas as pd



def metric():
    return int(random.random() * 1000)


def session_long():
    return random.choice('abcd')


DATA = [
    dict(sl=session_long(), metric=metric()),
    dict(sl=session_long(), metric=metric()),
    dict(sl=session_long(), metric=metric()),
    dict(sl=session_long(), metric=metric()),
    dict(sl=session_long(), metric=metric()),
    dict(sl=session_long(), metric=metric()),
    dict(sl=session_long(), metric=metric()),
    dict(sl=session_long(), metric=metric()),
    dict(sl=session_long(), metric=metric()),
    dict(sl=session_long(), metric=metric()),
    dict(sl=session_long(), metric=metric()),
    dict(sl=session_long(), metric=metric()),
    dict(sl=session_long(), metric=metric()),
    dict(sl=session_long(), metric=metric()),
    dict(sl=session_long(), metric=metric()),
    dict(sl=session_long(), metric=metric()),
    dict(sl=session_long(), metric=metric()),
    dict(sl=session_long(), metric=metric()),
    dict(sl=session_long(), metric=metric()),
    dict(sl=session_long(), metric=metric()),
    dict(sl=session_long(), metric=metric()),
    dict(sl='a', metric=None)  # IMPORTANT: Nan are excluded
]

df = pd.DataFrame(DATA)
df_agg = df.groupby(['sl'])['metric'].max().reset_index()


if __name__ == '__main__':
    print(df)
    print(df_agg)