Pandas.stack

https://kimdingko-world.tistory.com/211:[Pandas] 데이터의 재구조화 stack & unstack

• stack : 위에서 아래로 (행방향)
• unstack : 왼쪽에서 오른쪽으로 (열방향)

import pandas as pd
import numpy as np
import seaborn as sns
titanic = sns.load_dataset('titanic')
titanic.head()
wideUnstack = titanic.groupby(['pclass', 'sex'])[['survived', 'age', 'fare']].mean()
wideUnstack


# stack 함수는 시리즈를 반환
longStack = wideUnstack.stack()  # (level=-1, dropna=True)  마지막꺼..
longStack
longStack[1]
longStack[1]['female']
longStack[1, 'female']
longStack[1]['female']['age']
longStack[1, 'female'][['survived','fare']]

# 3개의 level   p/s/3
longStack.unstack(-1)  # (level=-1),(2)  2번index인 3이 열로가고 p/s 기준  
longStack.unstack(0)   # 0번 index인 p가 열로 가고  s/3 기준
longStack.unstack(1)   # 1번 index인 s가 열로 가고  p/3 기준 

# wideUnstack == longStack.unstack() 
longStack.unstack().equals(wideUnstack)  # True


wideUnstack
sns.heatmap(wideUnstack)