数据框列的选择性移动

Question

I have a dataframe with a number of columns and would selectively like to lag all bar one column (named target) by a specified number of steps.

def: shift_target(df,target,lag):
     df[df.columns != target].shift(lag)
     df = df.dropna()
     return df 

The above does not seem to work. Can i specify do not shift one column?

eg

Answer 1

You can construct a new dataframe. This is usually more efficient than manipulating an existing dataframe via Pandas methods. For example:

n = 4

res = df.iloc[:-n, 1:].copy()
res.index = df.index[n:]
res['Target'] = df['Target'].iloc[n:].values