How to change cell values of some specific rows without using for loop? python pandas-CodePudding

I have a dataframe df like this:

       symbol       P        Q          V   ...                                                 
0       00001     10.9      100         100     ...
1       00001     11.0      100         200     ... 
2       00001     11.0      100         200     ...     
3       00002     12.1     -100        -100     ...    
...       ...       ...         ...                                                       
46      00004     6.9       300         300     ...
47      00004     7.0      -200         100     ...    
48      00004     8.8       100         200     ...    
49      00004     8.8       100         200     ...

I also have an index list idx=[2, ... 49]. I would like to change the value of rows which indexings are in this list. For instance, I want to set

df.loc[i,'Q'] = df.loc[i-1,'V']

However, this kind of command threw some errors because it requires a specific index instead of an index list. I tried to use a for loop like this:

for i in idx:
    df.at[i, "Q"] = df.at[i-1, "V"]

Although it works, the speed would be slow down due to the for loop if the dataset is large. Is there any pandas command that could assign new values to a column of multiple rows based on an index list, like above?

CodePudding user response：

If you convert idx to a numpy array, it could work:

idx = [2, 3, 48, 49]
idx = np.array(idx)

# Don't forget .values to ignore index
df.loc[idx, 'Q'] = df.loc[idx-1, 'V'].values

Output:

>>> df
    symbol     P    Q    V
0        1  10.9  100  100
1        1  11.0  100  200
2        1  11.0  200  200
3        2  12.1  200 -100
46       4   6.9  300  300
47       4   7.0 -200  100
48       4   8.8  100  200
49       4   8.8  200  200