[FIXED] Why is this conditional statement in R dplyr generating an error message when it is generating the correct output and there are no NAs?

Issue

Please refer to Mikko Marttila answer below where he highlights the core issue with a better example. You don't need to waste your time going through all this OP gibberish.

I am working on a function with a for-loop and have broken it down into steps, as this my first ever for-loop.

The first section of code below, shown at the very bottom, generates a data frame called nCode and it is fine and produces no errors (leave the for-loop at i in 1:1 !!, just run the code without changes).

But when I run this second bit of code simulating the beginning of the 2nd loop run on the nCode data frame, it outputs fine but I get the error message "Problem with mutate() column concat_2. i concat_2 = ifelse(...). i NAs introduced by coercion". I can't see what's wrong with the ifelse(), it looks legit to me. Here's that second bit of code (to run after the first section of code is run):

i = 2
reSeq_prior <- str_c("reSeq_",i-1)
concat_col <- str_c("concat_",i)

nCode <- if(i==1){
  nCode %>% mutate(!! concat_col:= as.numeric(paste0(seqBase,".",grpRnk)))} else {
    nCode %>% mutate(!! concat_col:= ifelse(
      !!rlang::sym(reSeq_prior)%%1 > 0, 
      !!rlang::sym(reSeq_prior),
      as.numeric(paste0(!!rlang::sym(reSeq_prior),".",grpRnk))
      )
    )
  }
nCode

Here's the good output I get when running these two sections of code (I am resisting my urge to use suppressWarnings(), I'd rather understand the problem):

> nCode
# A tibble: 15 x 10
   Name  Group nmCnt seqBase grpRnk concat_1 alloc_1 merge_1 reSeq_1 concat_2
   <chr> <dbl> <int>   <int>  <dbl>    <dbl>   <dbl>   <dbl>   <dbl>    <dbl>
 1 R         0     1       1      0      1       1       1       1        1  
 2 R         0     2       2      0      2       2.1     2.1     2.1      2.1
 3 X         0     1       1      0      1       1       1       1        1  
 4 X         1     2       2      1      2.1     2.1     2.1     2.1      2.1
 5 X         1     3       2      2      2.2     2.2     2.2     2.2      2.2
 6 X         0     4       3      0      3      NA       3       3        3  
 7 X         0     5       4      0      4      NA       4       4        4  
 8 X         0     6       5      0      5      NA       5       5        5  
 9 B         0     1       1      0      1       1       1       1        1  
10 R         0     3       3      0      3       2.2     2.2     2.2      2.2
11 R         2     4       4      1      4.1    NA       4       3        3.1
12 R         2     5       4      2      4.2    NA       4       3        3.2
13 X         3     7       6      1      6.1    NA       6       6        6.1
14 X         3     8       6      2      6.2    NA       6       6        6.2
15 X         3     9       6      3      6.3    NA       6       6        6.3

First section of code:

library(dplyr)
library(stringr)

myDF1 <-
  data.frame(
    Name = c("R","R","X","X","X","X","X","X","B","R","R","R","X","X","X"),
    Group = c(0,0,0,1,1,0,0,0,0,0,2,2,3,3,3)
  )

nCode <-  myDF1 %>%
  group_by(Name) %>%
  mutate(nmCnt = row_number()) %>%
  ungroup() %>%
  mutate(seqBase = ifelse(Group == 0 | Group != lag(Group), nmCnt,0)) %>%
  mutate(seqBase = na_if(seqBase, 0)) %>%
  group_by(Name) %>%
  fill(seqBase) %>%
  mutate(seqBase = match(seqBase, unique(seqBase))) %>%
  ungroup %>%
  mutate(grpRnk = ifelse(Group > 0, sapply(1:n(), function(x) sum(Name[1:x]==Name[x] & Group[1:x] == Group[x])),0))
  
loopCntr <- nrow(unique(myDF1[myDF1$Group!=0,]))
  
for(i in 1:1) {
  
  reSeq_prior <- str_c("reSeq_",i-1)
    
  concat_col <- str_c("concat_",i)

  nCode <- if(i==1){
    nCode %>% mutate(!! concat_col:= as.numeric(paste0(seqBase,".",grpRnk)))} else {
      nCode %>% mutate(!! concat_col:= as.numeric(paste0(!!rlang::sym(reSeq_prior),".",grpRnk)))
    }
  
    index <- filter(nCode, Group !=0) %>%
      select(all_of(concat_col)) %>%
      distinct() %>%
      mutate(truncInd = trunc(get(concat_col))) %>%
      group_by(truncInd) %>%
      mutate(cumGrp = cur_group_id()) %>%
      ungroup() %>%
      select(-truncInd,cumGrp,concat_col)
  
  # below inserts a 1 in 1st row of index if lowest element count group is >= 2  
    index <- if(ifelse(loopCntr > 0, min(index[[concat_col]]), Inf) >= 2){
      tmp <- data.frame(cumGrp = 1, concat = 1)
      names(tmp)[2] <- concat_col
      rbind(tmp,index)
    }else{index}
    
    nCode <- nCode %>%
      mutate(alloc = index[[concat_col]][index$cumGrp==1][nmCnt]) %>%
      mutate(merge = ifelse(is.na(alloc),seqBase,alloc)) %>%
      group_by(Name) %>%
      mutate(reSeq = match(trunc(merge), unique(trunc(merge)))) %>%
      mutate(reSeq = (reSeq + round(merge%%1 * 10,0)/10)) %>%
      ungroup()
    
    nCode <- nCode %>%
      rename_with(~ str_c(.x, "_", i), c("alloc", "merge", "reSeq"))
   
  }

Solution

Both branches are evaluated in ifelse(). The warnings are generated from NAs in the no branch, even though the final result will include the value from the yes branch.

Here’s a simplified example:

a <- c(1.1, 2.1, 1)
b <- c(0, 2, 1)

ifelse(a != trunc(a), a, as.numeric(paste0(a, ".", b)))
#> Warning in ifelse(a != trunc(a), a, as.numeric(paste0(a, ".", b))): NAs
#> introduced by coercion
#> [1] 1.1 2.1 1.1

Which is essentially equivalent to:

y <- a
n <- as.numeric(paste0(a, ".", b))
#> Warning: NAs introduced by coercion

ifelse(a != trunc(a), y, n)
#> [1] 1.1 2.1 1.1

To avoid the warning, write code that won’t generate warnings in either branch:

ifelse(a != trunc(a), a, a + b / 10)
#> [1] 1.1 2.1 1.1

Answered By - Mikko Marttila

Answer Checked By - Mildred Charles (PHPFixing Admin)