简化 pandas groupby().agg() 的代码

Question

I have a dataframe that I'm trying to group by and get sum for multiple columns, for which I have below code:

df=df.groupby(['year','month']).agg({'A':['sum'],'B':['sum'],'C':['sum'],'D':['sum']})

Is there a way I could change the arguments in agg() to iterate through a list? I'm trying something like this, but obviously it's not working.

col=['A','B','C','D']
df=df.groupby(['year','month']).agg({c for c in col})

Thank you very much!

Answer 1

You are very close. Note you are passing into agg() a set, not a dictionary. A dictionary has a pair of key: value where you just have a value .

df=df.groupby(['year','month']).agg({c: ['sum'] for c in df.columns})

Because:

{c: ['sum'] for c in df.columns}
>>> {'A':['sum'],'B':['sum'],'C':['sum'],'D':['sum']}

In contrast to what you wrote:

{c for c in df.columns}  # you iterated over cols, probably forgot cols=df.columns before. Changed it to df.columns here
>>> {'A', 'B', 'C', 'D'}

Edit: I'm also assuming you are not interested in summing all your columns, only 'A' to 'D'. If it is in fact your intention, like stated in other comments under your question, you can just do:

df.groupby(['year','month']).sum()

Or

df.groupby(['year','month']).agg('sum')