pandas 行列转换

# -*- coding:utf-8 -*-
import pandas as pd

# 一行转多行
df = pd.DataFrame(
    {'Code': ['1111', '2222'],
     'Color': ['a, b, c,d', 'blue, pink, yellow'],
     'Size': ['1, 2, 3', '7, 6, 10']}
)
id_df = df[["Code"]]

# stack函数会将数据从”表格结构“变成”花括号结构“，即将其行索引变成列索引
# 参考 https://www.cnblogs.com/bambipai/p/7658311.html
for col in df.columns[1:]:
    split_df = df[col].str.split(",", expand=True).stack().reset_index(level=1, drop=True).rename(col)
    id_df = id_df.join(split_df)

# 删除行索引并重建
convert_df = id_df.reset_index(drop=True)

print(convert_df)

df = pd.DataFrame(
    {"id": [2, 3, 3, 3],
     "name": ["tx", "hw", "hw", "hw"],
     "sub": ["game", "mobile", "cell", "cloud"]
     }
)


# 多行转一行

def collect(ls):
    return ",".join(ls)


df2 = df.groupby(["id", "name"])["sub"].apply(collect).rename("subs")
df2.reset_index()