Selecting the first two rows based on condition

I have a database like this :

And i want my output to select the first two rows of each ID found and place the earliest value as new value and the oldest one as old value :

Thanks a lot for your help!

>Solution :

This is a variation on a pivot with filtering of the data:

out = (df
   .assign(col=lambda d: d.groupby('ID').cumcount())  # enumerate values
   .query('col < 2') # select top 2 per group
   .pivot(index='ID', columns='col', values='Value')  # reshape
   .set_axis(['New Value', 'Old Value'], axis=1).reset_index() # clean-up
)

Output:

      ID  New Value  Old Value
0  12345     123.78     556.87
1  56789      44.00     434.20