How to get the difference between first row and current row in pandas

Question

I have 2 columns in dataframe (date & sell_price). My expected output is like this. I want to add one more column named profit in dataframe that needs to be calculated by current sell_price - first sell price (stars one)

        date            sell_price  profit(needs to be added)
    0   2018-10-26       **21.20**    NaN
    1   2018-10-29       15.15      -6.05
    2   2018-10-30       15.65      -5.55
    3   2018-10-31        0.15     -21.05
    4   2018-11-01        5.20     -16.00

I know the diff in pandas that gives difference between consecutive rows. How can we achieve expected o/p with diff or any other function on pandas?

Answer 1

For general Index like DatetimeIndex use iloc with iat, but it working only with positions, so necessary get_loc:

pos = df.columns.get_loc('sell_price')
df['profit'] =  df.iloc[1:, pos] - df.iat[0, pos]

If default RangeIndex use loc with at:

df['profit'] =  df.loc[1:, 'sell_price'] - df.at[0, 'sell_price']

---

print (df)
         date  sell_price  profit
0  2018-10-26       21.20     NaN
1  2018-10-29       15.15   -6.05
2  2018-10-30       15.65   -5.55
3  2018-10-31        0.15  -21.05
4  2018-11-01        5.20  -16.00