Group many columns by index instead of by label in a pandas Groupby statement

Question

I have a dataframe with +60 columns. I need to group by two columns, and only need one column values summed. The problem is if I don't manually type the names of all the column labels in the groupby statement, columns not included will not appear in the output.

Instead of something like this:

df_final.groupby(by=['OrderNo','ItemSKU','CustName',.......'20th Column'],as_index=False).sum()

I'd like to do something like this:

df_final.groupby(by=[:20],as_index=False).sum()

How can I do this and avoid typing all those column names?

Here is a print of the column datatypes:

>>> print(df_final.dtypes)
OrderNo                   float64
PledgeID                    int64
ReferrerID                float64
FulfillmentStatus          object
FundingDate                object
PaymentMethod             float64
Appearance                 object
Name                       object
Email                      object
Amount                     object
PlatformFee                object
PerkID                    float64
Perk                       object
ShippingName               object
ShippingPhoneNumber        object
ShippingAddress            object
ShippingAddress2           object
ShippingCity               object
ShippingState/Province     object
ShippingZip/PostalCode     object
ShippingCountry            object
ItemSKU                    object
ArticleName                object
UPC                       float64
ArticleQty                  int64
dtype: object
>>> 

Answer 1

您可以将前20列名称转换为列表：

df_final.groupby(by=df_final.columns[:20].tolist(),as_index=False).sum()