减去数据帧中的两个值

Question

I have a dataframe (df) that has 2 columns by the name [1,2], I used the following code to add 2 more columns: largest and second largest that display at each row which is the highest number and the second highest.

    df = RSRP_per_sec(2)
    df2=df.copy()
    df2['highest']=0
    df2['second_highest']=0
    for i in range(len(df)):
        l=sorted([(x, df.iloc[i][x]) for x in df.columns], key=lambda y: y[1])
        df2['highest'].iloc[i]=l[-1][0]
        df2['second_highest'].iloc[i]=l[-2][0]

and now df dataframe is like this where the columns are ['1','2','highest','second_highest'] and displaying the first row:

1        2       highest    second_highest

-86     -102      1           2

Now I would like to add one more column that calculates the difference between the highest and second highest by checking at each row which column is the highest and which is the second and subtract them from each others.

Note: for my case now its easy as there are only 2 columns to subtract but I would like to expand the number of columns that highest and second highest columns will have a different values.

Answer 1

Take the column values and use them to index the corresponding columns:

df["diff_highest"] = df[str(df["highest"].values[0])] - df[str(df["second_highest"].values[0])]

Btw you can also get the two highest values by using .sort_values() to each row:

df["1"] = [-86]
df["2"] = [-102]
df["3"] = [12]
df["highest"] = [None]
df["second_highest"] = [None]

df.iloc[0].sort_values(ascending=False).dropna().index[:2]

outputs

Index(['3', '1'], dtype='object')

I would suggest to set "highest" and "second_highest" to None so it does not conflict the sorting.