ifelse replace value if it is lower than previous - r

Question

I am working with a dataset which has some errors in the data. Numbers are sometimes registered wrong. Here is some toy data example:
The issue is that the Reversal column should only be counting up (per unique ID). So in a vector of 0,0,0,1,1,1,0,1,2,2,0,0,2,3, the 0's following the 1 and 2 should not be 0's. Instead, they should be equal to whatever value came before. I tried to remedy this by using the lag function from the dplyr package:
Data$Reversal <- ifelse(Data$Reversal < lag(Data$Reversal), lag(Data$Reversal), Data$Reversal) .
But this results in numerous issues:
The first value becomes NA. I've tried using the default=Data$Reversal call in the lag function but to no avail.
The Reversal value should reset to 0 for each Unique ID. Now it continues across ID's. I tried a messy code using group_by(ID) but could not get this to work, as it broke my earlier ifelse function.
This only works when there is 1 error. But if there are two errors in a row it only fixes 1 value.
Alternatively, I found this thread in which the answer provided by Andrie also seems promising. This fixes problem 1 and 3, but I can't get this code to work per ID (using the group_by function).
Andrie's answer:
local({
r <- rle(data)
x <- r$values
x0 <- which(x==0) # index positions of zeroes
xt <- x[x0-1]==x[x0+1] # zeroes surrounded by same value
r$values[x0[xt]] <- x[x0[xt]-1] # substitute with surrounding value
inverse.rle(r)
})
Any help would be much appreciated.

Answer 1

I think cummax does exactly what you need.
Base R
dat$Reversal <- ave(dat$Reversal, dat$ID, FUN = cummax)
dat
# ID Owner Reversal Success
# 1 1 A 0 0
# 2 1 A 0 0
# 3 1 A 0 0
# 4 1 B 1 1
# 5 1 B 1 0
# 6 1 B 1 0
# 7 1 error 1 0
# 8 1 error 1 0
# 9 1 B 1 0
# 10 1 B 1 0
# 11 1 C 1 1
# 12 1 C 2 0
# 13 1 error 2 0
# 14 1 C 2 0
# 15 1 C 3 1
# 16 2 J 0 0
# 17 2 J 0 0
dplyr
dat %>%
group_by(ID) %>%
mutate(Reversal = cummax(Reversal)) %>%
ungroup()
data.table
as.data.table(dat)[, Reversal := cummax(Reversal), by = .(ID)][]
Data, courtesy of https://extracttable.com/
dat <- read.table(header = TRUE, text = "
ID Owner Reversal Success
1 A 0 0
1 A 0 0
1 A 0 0
1 B 1 1
1 B 1 0
1 B 1 0
1 error 0 0
1 error 0 0
1 B 1 0
1 B 1 0
1 C 1 1
1 C 2 0
1 error 0 0
1 C 2 0
1 C 3 1
2 J 0 0
2 J 0 0")