How to replace NAs based on preceding and following numbers are zero

Question

I have a large matrix full of NAs. For each row I would like to replace the NAs with zeros but only if the preceding and following values are zero. The number of NAs between zeros can vary so in the small example below NAs rows a.1 and a.1.2 will be replaced but other rows will remain NA

a.1<-c(1,1,1,0,NA,NA,0,1)
a.1.2 <- c(1,1,0,NA,NA,NA,0,0)
a.2<-c(1,0,1,1,NA,NA,0,1)
a.3<-c(1,0,0,0,NA,NA,1,0)
a.4<-c(1,0,0,1,NA,NA,1,0)

a_matrix<-as.matrix(rbind(a.1,a.1.2,a.2,a.3,a.4))

Answer 1

We can do it with a custom function:

my_func <- function(x) {
  r <- rle(is.na(x))
  vals <- r$values
  lengths <- r$lengths
  
  for (i in 2:(length(vals) - 1)) {
    if (vals[i] & !vals[i-1] & !vals[i+1] & x[sum(lengths[1:(i-1)])] == 0 & x[sum(lengths[1:i]) + 1] == 0) {
      x[(sum(lengths[1:(i-1)])+1):(sum(lengths[1:i]))] <- 0
    }
  }
  return(x)
}

t(apply(a_matrix, 1, my_func))

      [,1] [,2] [,3] [,4] [,5] [,6] [,7] [,8]
a.1      1    1    1    0    0    0    0    1
a.1.2    1    1    0    0    0    0    0    0
a.2      1    0    1    1   NA   NA    0    1
a.3      1    0    0    0   NA   NA    1    0
a.4      1    0    0    1   NA   NA    1    0