Apply shift function group of rows in pandas

Question

I am trying to create new column feature from curr_value feature. Here, I want get the value from previous current value, using df.curr_value.shift(fill_value=-1) I can get the previous value. How do apply shift to only unique 'uid' columns. For example, In below example, I want to apply shift to uniqie ID values. df[df['uid']==1].curr_value.shift(fill_value=-1) Is there any other way in pandas so that I dont need to iterate unique uid values?

import pandas as pd
import numpy as np
events  = np.array([[1, 1, 1], 
[1, 2, 4], 
[1, 3, 3], 
[1, 3, 4], 
[2, 1, 4], 
[2, 2, 3]])
df=pd.DataFrame(data=events, columns=["uid", "eid", "curr_value"])

Answer 1

Yes we can do it with groupby

df['pre'] = df.groupby('uid')['curr_value'].shift(fill_value=-1)
Out[5]: 
0   -1
1    1
2    4
3    3
4   -1
5    4
Name: curr_value, dtype: int32