熊猫据帧字段转换为键和值蟒蛇

Question

Date   sector weight 
0 2011-01-10  a  12 
1 2011-01-10  b  23 
2 2011-01-11  c  34 
3 2011-01-11  d  54 

• 我DF这样的，我想这个结果在下面的格式

  {2011-01-10：{A：12，B：23 }，2011-01-11：{C：34，d：54}}

Answer 1

使用groupby + apply + to_json：

#if necessary convert to string 
df['Date'] = df['Date'].astype(str) 
df = df.groupby('Date').apply(lambda x: dict(zip(x['sector'], x['weight']))).to_json() 
print (df) 
{"2011-01-10":{"b":23,"a":12},"2011-01-11":{"c":34,"d":54}} 

Answer 2

grpby = df.set_index(['Date', 'sector']).weight.groupby('Date') 
{n: g.xs(n).to_dict() for n, g in grpby} 

{'2011-01-10': {'a': 12, 'b': 23}, '2011-01-11': {'c': 34, 'd': 54}}