使用Pandas转换为长面板数据格式 [英] Converting to long panel data format with pandas

问题描述

我有一个DataFrame，其中的行代表时间，而列则代表个人.我想以一种有效的方式将其转换为大熊猫的长面板数据格式，因为DataFame相当大.我想避免循环.这是一个示例:以下DataFrame:

I have a DataFrame where rows represent time and columns represent individuals. I want to turn it into into long panel data format in pandas in an efficient manner, as the DataFames are rather large. I would like to avoid looping. Here is an example: The following DataFrame:

      id    1    2
date              
20150520  3.0  4.0
20150521  5.0  6.0

应转换为:

date        id        value
20150520    1         3.0
20150520    2         4.0
20150520    1         5.0
20150520    2         6.0

由于数据大小，速度对我而言真正重要.如果要权衡，我宁愿选择它而不是优雅.尽管我怀疑我可能缺少一个相当简单的功能，但熊猫应该能够处理该功能.有什么建议吗?

Speed is what's really important to me, due to the data size. I prefer it over elegance if there is a tradeoff. Although I suspect I mam missing a rather simple function, pandas should be able to handle that. Any suggestions?

推荐答案

我认为您需要

I think you need stack with reset_index:

print (df)
            1    2
date              
20150520  3.0  4.0
20150521  5.0  6.0

df = df.stack().reset_index()
df.columns = ['date','id','value']
print (df)
       date id  value
0  20150520  1    3.0
1  20150520  2    4.0
2  20150521  1    5.0
3  20150521  2    6.0

---

print (df)
id          1    2
date              
20150520  3.0  4.0
20150521  5.0  6.0

df = df.stack().reset_index(name='value')
print (df)
       date id  value
0  20150520  1    3.0
1  20150520  2    4.0
2  20150521  1    5.0
3  20150521  2    6.0

这篇关于使用Pandas转换为长面板数据格式的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！