在pandas中比较行的子集。

huangapple go评论71阅读模式
英文:

Comparing subset of rows in pandas

问题

我在想是否有一种好的方法来比较pandas中的一部分行?
假设我有一个带有以下内容的df:

id in_test value
1 True 5
2 True 5
1 False 7
2 False 8

我想要的结果是id和从in_test从true到false的差异(或百分比变化)的df。

我知道我可以将表旋转然后执行逐行计算,或者创建一个过滤后的df并将其与另一个过滤后的df合并,然后逐行计算。

我在想是否有一种在一行中完成此操作的python方法?可能使用pandas函数?

百分比差异的输出将是:

id value
1 +40%
2 +60%

差异的输出将是:

id value
1 2
2 3
英文:

I was wondering if there is nice way to compare a subset of rows in pandas?
let's say I have a df with:

id in_test value
1 True 5
2 True 5
1 False 7
2 False 8

I would like the resulting df with id and difference (or percentage change) from in_test from true to false.

I know I could pivot the table and then perform row wise calculations, or create a filtered df and merge it with another filtered df and then compute it row wise.

I was wondering if there is python way of doing this in one line? With probably a pandas function?

The output for percentage diff would be :

id value
1 +40%
2 +60%

The output for diff would be :

id value
1 2
2 3

(or minus -2 & -3 i guess I would have top define some kind of order) )

答案1

得分: 1

按id分组,计算value的差异,然后重置索引

差异

按id分组,计算value的百分比变化,然后重置索引

百分比变化

英文:
df.groupby('id')['value'].apply(lambda x: x.diff().values[1]).reset_index()

difference

df.groupby('id')['value'].apply(lambda x: x.pct_change().values[1] * 100).reset_index()

percentage difference

huangapple
  • 本文由 发表于 2023年3月21日 01:11:56
  • 转载请务必保留本文链接:https://go.coder-hub.com/75793312.html
匿名

发表评论

匿名网友

:?: :razz: :sad: :evil: :!: :smile: :oops: :grin: :eek: :shock: :???: :cool: :lol: :mad: :twisted: :roll: :wink: :idea: :arrow: :neutral: :cry: :mrgreen:

确定