Conditional imputation with LOCF

I've this example of longitudinal data. I need to impute 0, 999 or -1 values according to what occurs before.

ID = c(1,1,1,1,1,2,2,2,3,3,3,4,4,4,5,5,5,5,6,6,6,6,6,6,6,6) 
Oxy = c(0, 999, 1, 999, 999, 0, 0, 999, 999, 0, 0, -1, 0, 999, 1, 1, -1, 1, 999, -1, 0, -1, 1,0, 999, 0) 
Y = c(2010,2011,2012,2013,2014,2011,2012,2013,2010,2011,2012,2010,2011,
      2012,2010,2011,2012,2013,2014,2015,2016,2017, 2018,2019,2020, 2021) 
Oxy2 = c(0, 999, 1, 1, 1, 0, 0, 999, 999, 0, 0, -1, 0, 999, 1, 1, 1, 1, 999, -1, 0, -1, 1, 1,1,1) 
df = data.frame(ID, Oxy, Y, Oxy2)

Basically, I'd like to get Oxy2 from Oxy. I need to keep 999 when previous values of Oxy are 0 or -1, and replace everything else comes after the first 1 appears, considering the group ID over time.

ID Oxy    Y Oxy2
  1   0 2010    0
  1 999 2011  999
  1   1 2012    1
  1 999 2013    1
  1 999 2014    1
  2   0 2011    0
  2   0 2012    0
  2 999 2013  999
  3 999 2010  999
  3   0 2011    0
  3   0 2012    0
  4  -1 2010   -1
  4   0 2011    0
  4 999 2012  999
  5   1 2010    1
  5   1 2011    1
  5  -1 2012    1
  5   1 2013    1
  6 999 2014  999
  6  -1 2015   -1
  6   0 2016    0
  6  -1 2017   -1
  6   1 2018    1
  6   0 2019    1
  6 999 2020    1
  6   0 2021    1

Thanks for your suggestion.