Using regex to set a specific digit to NA?

Question

Sample of df:

   LASSO_deviance LASSO_AUC
68          0.999     0.999
2           1.000     1.000
39          1.000     1.005
7            1.02       1.2

I want to set cells which contain 1.000 to either NA or 0, in preferential order.

I've tried something like: df %>% mutate_at(vars(LASSO_deviance, LASSO_AUC), funs(gsub(pattern = "1{1}[^.{1,}]", 0, x = .))) with no luck.

@TimBiegeleisen numeric, but I've tried setting the cols to text but both don't work. — user126082, Sep 26 '18 at 10:22
`1.000 to either NA or 0` **WHAT DO YOU WANT?** Replace `1 -> NA` or `1 -> 0`. — Andre Elrico, Sep 26 '18 at 10:23
btw: avoid using regex when you're working with numeric-only input. — Andre Elrico, Sep 26 '18 at 10:29

score 0 · Accepted Answer · answered Sep 26 '18 at 10:25

tt <- "LASSO_deviance LASSO_AUC
68          0.999     0.999
2           1.000     1.000
39          1.000     1.005
7            1.02       1.2"

dat <- read.table(text = tt, header = T)

No need for regex because you can simply find where your data is equal to 1.000

dat[dat == 1.000] <- NA # or dat[dat == 1.000] <- 0
dat

# LASSO_deviance LASSO_AUC
# 68          0.999     0.999
# 2              NA        NA
# 39             NA     1.005
# 7           1.020     1.200

Using regex to set a specific digit to NA?

1 Answers1