[英]Pandas create dataframe from dict converting tuple to index
有没有方便的方法实现一个函数make_dataframe
,使用如下
mydict = {
('tom', 'gray') : [1,2,3,4,5],
('bill', 'ginger') : [6,7,8,9,10],
}
make_dataframe(mydict, tupleLabels=['catname', 'catcolor'], valueLabel='weight')
预期结果
| catname | catcolor | weight |
| tom | gray | 1 |
| tom | gray | 2 |
| tom | gray | 3 |
| tom | gray | 4 |
| tom | gray | 5 |
| bill | ginger | 6 |
| bill | ginger | 7 |
| bill | ginger | 8 |
| bill | ginger | 9 |
| bill | ginger | 10 |
听起来不太难,我只是不想重新发明轮子
您可以使用数据帧创建自己的功能unstack
重命名使用的标签后rename_axis
:
def make_dataframe(dictionary , tupleLabels , valueLabel):
return (pd.DataFrame(dictionary).rename_axis(tupleLabels,axis=1)
.unstack().reset_index(tupleLabels,name=valueLabel))
out = make_dataframe(mydict, tupleLabels=['catname', 'catcolor'], valueLabel='weight')
print(out)
catname catcolor weight
0 tom gray 1
1 tom gray 2
2 tom gray 3
3 tom gray 4
4 tom gray 5
0 bill ginger 6
1 bill ginger 7
2 bill ginger 8
3 bill ginger 9
4 bill ginger 10
您的字典格式错误,无法轻松转换为 Pandas DataFrame。
我建议执行以下操作:
mydict = {
('tom', 'gray') : [1,2,3,4,5],
('bill', 'ginger') : [6,7,8,9,10],
}
l = [ [ k[0], k[1], val ] for k, v in mydict.items() for val in v ]
df = pd.DataFrame(l, columns=['catname', 'catcolor', 'weight'])
其中产生:
catname catcolor weight
0 tom gray 1
1 tom gray 2
2 tom gray 3
3 tom gray 4
4 tom gray 5
5 bill ginger 6
6 bill ginger 7
7 bill ginger 8
8 bill ginger 9
9 bill ginger 10
声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.