如何在 df 与其他 df 连接的列中执行函数 sum()

Question

I am using concat to merge 5 equal df into one and get the total sum() of cost.

These values are not real, just an example of what df looks like

What I tried:

import pandas as pd

g = {"id": "1515", "cost": "100"}
b = {"id": "1515", "cost": "100"}
f = {"id": "1515", "cost": "100"}
c = {"id": "1515", "cost": "100"}
o = {"id": "1515", "cost": "100"}

all_vendors = pd.concat([g, b, f, c, o])

Data types

all_vendors.dtypes

Campaign          object
campaignid       float64
Campaign_name     object
Cost              object
Month             object
Year & month      object
dtype: object

Attempts

Attempt #1:

all_vendors.Cost.sum()

Results in:

TypeError: can only concatenate str (not "float") to str

Attempt #2:

all_vendors.Cost.astype(str)
all_vendors.Cost.sum()

Results in:

TypeError: can only concatenate str (not "float") to str

Attempt #3:

all_vendors.Cost.astype(float)
all_vendors.Cost.sum()

Results in:

ValueError: could not convert string to float: '100'

Answer 1

Your problem is that you're not reassigning your astype call back to your DataFrame :

import pandas as pd

data = {
  "id": ['1,515','1,515','1,515','1,515','1,515'],
  "cost": ['1,000','1,000','1,000','1,000','1,000']
}

all_vendors = pd.DataFrame.from_dict(data)

all_vendors['cost'] = all_vendors.cost.str.replace(',','').astype(float)
print(all_vendors.cost.sum())
# Output: 500

As mentioned in the comments, use str.replace to remove any commas you have in your strings

Answer 2

you first need to convert dataframe to float to be able to add numbers with decimals, for them you use DataFrame.astype

import pandas as pd
g = pd.DataFrame({"id": ["1515"], "cost": ["100"]})
b = pd.DataFrame({"id": ["1515"], "cost": ["100"]})
f = pd.DataFrame({"id": ["1515"], "cost": ["100"]})
c = pd.DataFrame({"id": ["1515"], "cost": ["100"]})
o = pd.DataFrame({"id": ["1515"], "cost": ["100"]})
all_vendors = pd.concat([g, b, f, c, o])

---

if you have ',' in your string, then you need:

all_vendors['cost']=all_vendors['cost'].str.replace(',','')

Then you calculate the sum:

all_vendors.astype(float).cost.sum()

Output:

500.0

if you want to work with the float type data frame you need to assign it:

all_vendors2=all_vendors.astype(float)
all_vendros2.cost.sum()

Output:

500

Answer 3

I got this to work on my end with a value of 500:

df_list = [pd.DataFrame(data={"id": ["1515"], "cost": ["100"]}) for i in range(5)]
pd.concat(df_list).cost.astype(float).sum()

So long as they're dataframes and you convert the strings to floats, it looks good.

Answer 4

Check if this helps.This will give total of id's.

import pandas as pd

g = pd.DataFrame({"id": ["1515"], "cost": ["100"]})
b = pd.DataFrame({"id": ["1515"], "cost": ["100"]})
f = pd.DataFrame({"id": ["1515"], "cost": ["100"]})
c = pd.DataFrame({"id": ["1515"], "cost": ["100"]})
o = pd.DataFrame({"id": ["1515"], "cost": ["100"]})
all_vendors = pd.concat([g, b, f, c, o])

a=pd.DataFrame.from_records(all_vendors).astype(float).groupby('id').sum().T.to_dict()
print(a)