To drop columns, you can use pandas' drop() function. It takes a single column as well as a list of columns, and in this example, additional optional arguments indicate which is the axis along which to drop and whether or not to drop columns in place:
df.drop(['col1','col2'], axis=1, inplace=True)
print(df)
The output is as follows:
col3 new_col1 new_col2 new_col3 new_col4 0 x 0 5 5 1 y 0 7 7 2 z 0 9 9