将数据框列表示为其他列的线性组合

Question

I have a data frame with 10 columns of data and 2000 rows. It also has additional columns that need to be ignored:

df1 = pd.DataFrame(np.random.randint(0,100,size=(2000, 10)), columns=list('ABCDEFGHIJ'))
df1['Company Name']=stringlist1

The names of columns can change from run to run as different files have different column names. The only thing common is that the data to be considered starts from the 7th column onwards, for the next 10 columns. I have a several lists, each containing 10 weights, some of which are zero, others are non-zero, adding up to 1. Example:

wt1=[0.0,0.34,0.05,0.0,0.1,0.01,0.0,0.0,0.5,0.0]

I need to define a new df1 column that is the linear combination of the 10 columns, with the weights specified in wt1.

How do I do that? Mind you, the names of columns (ABCD...) cannot appear in the summation expression, as the above code needs to be reusable for data where column names can be different (they are being read in from an Excel sheet).

I tried:

icollist1=[icol1 for icol1,val1 in enumerate(wt1) if val1>0.0]
for icol1 in icollist1:
    df1['Weighted Sum']+=np.asarray(wt1[icol1])*df1[colnames1[icol1]]

where colnames1 is a list of columns extracted from the Excel file this dataframe was read from.

I get errors:

TypeError: can't multiply sequence by non-int of type 'float'
...
During handling of the above exception, another exception occurred:
...
TypeError: can't multiply sequence by non-int of type 'float'

Answer 1

Try this for the example you provide

df1 = pd.DataFrame(np.random.randint(0,100,size=(2000, 10)), columns=list('ABCDEFGHIJ'))
wt1=[0.0,0.34,0.05,0.0,0.1,0.01,0.0,0.0,0.5,0.0]

df1.mul(wt1, axis=1).sum(axis=1)

In case you have more than 10 columns and you want to multiple from 7th columns onward:

df1 = pd.DataFrame(np.random.randint(0,100,size=(2000, 20)))
wt1=[0.0,0.34,0.05,0.0,0.1,0.01,0.0,0.0,0.5,0.0]
df1.iloc[:,6:16].mul(wt1, axis=1).sum(axis=1)