CODEWITHSUNDEEP
rdataframestringr
user8959427
user8959427

Reputation: 2067

replacing NA values using stringr

My data looks as follows:

   var1 var2 var3
1    9V  .6V  77V
2    6V  .3V  15V
3    9V  .7V 114V
4   12V 1.0V 199V
5   14V 1.2V 245V
6   13V  .8V 158V
7   11V  .6V 136V
8   11V  .7V 132V
9   12V  .8V 171V
10  11V  .7V 155V
11  13V  .8V 166V
12  11V  .7V 138V
13  11V  .9V 173V
14   9V  .8V 143V
15   8V  .7V 105V
16   4V  .4V  21V
17   8V  .4V  26V
18  10V  .8V 154V
19   9V  .8V 130V
20  10V  .7V 113V
21  10V  .6V 102V
22  11V  .8V 135V
23   9V  .7V 120V
24  10V  .7V 126V
25   7N  .6N 124N
26  14N 1.1N 210N

The last 2 rows contain N. I am trying to set to NA these observations which contain N.

I am trying some combination of str_detect and str_replace but I cannot seem to get it working.

Additionally I have other (very rarely) letters, i.e. M and P - I would like to find a way to set if the observations contains one of these letters, then set that observation to NA. i.e. c(var1:var3) %in% str_detect(c("N", "M", "P"))... str_replace_all.

Data:

    structure(list(var1 = c("9V", "6V", "9V", "12V", "14V", "13V", 
"11V", "11V", "12V", "11V", "13V", "11V", "11V", "9V", "8V", 
"4V", "8V", "10V", "9V", "10V", "10V", "11V", "9V", "10V", "7N", 
"14N", "7V", "5V", "7V", "9V", "8V", "8V", "5V", "4V", "4V", 
"5V", "7V", "5V", "6V", "8V", "9V", "6V", "6V", "7V", "8V", "7V", 
"8V", "8V", "7V", "8V"), var2 = c(".6V", ".3V", ".7V", "1.0V", 
"1.2V", ".8V", ".6V", ".7V", ".8V", ".7V", ".8V", ".7V", ".9V", 
".8V", ".7V", ".4V", ".4V", ".8V", ".8V", ".7V", ".6V", ".8V", 
".7V", ".7V", ".6N", "1.1N", ".4V", ".3V", ".4V", ".6V", ".5V", 
".6V", ".4V", ".3V", ".2V", ".3V", ".4V", ".3V", ".3V", ".5V", 
".6V", ".4V", ".4V", ".4V", ".5V", ".4V", ".4V", ".5V", ".4V", 
".4V"), var3 = c("77V", "15V", "114V", "199V", "245V", "158V", 
"136V", "132V", "171V", "155V", "166V", "138V", "173V", "143V", 
"105V", "21V", "26V", "154V", "130V", "113V", "102V", "135V", 
"120V", "126V", "124N", "210N", "35V", "9V", "48V", "91V", "81V", 
"80V", "14V", "11V", "7V", "13V", "34V", "18V", "15V", "58V", 
"76V", "29V", "30V", "31V", "32V", "34V", "57V", "58V", "52V", 
"49V")), row.names = c(NA, 50L), class = "data.frame")

Upvotes: 2

Views: 402

Answers (3)

M--
M--

Reputation: 28825

The dplyr-stringr solution that you were trying to figure out would be like below:

library(stringr)
library(dplyr)

df1 %>% 
  mutate_at(vars(var1:var3), 
            list(~str_replace_all(., "N$|M$|P$", replacement = NA_character_)))

#>    var1 var2 var3
#> 1    9V  .6V  77V
#> 2    6V  .3V  15V
#> 3    9V  .7V 114V
#> 4   12V 1.0V 199V
#> 5   14V 1.2V 245V
## ...
#> 20  10V  .7V 113V
#> 21  10V  .6V 102V
#> 22  11V  .8V 135V
#> 23   9V  .7V 120V
#> 24  10V  .7V 126V
#> 25 <NA> <NA> <NA>
#> 26 <NA> <NA> <NA>
#> 27   7V  .4V  35V
#> 28   5V  .3V   9V
#> 29   7V  .4V  48V
#> 30   9V  .6V  91V
## ...
#> 45   8V  .5V  32V
#> 46   7V  .4V  34V
#> 47   8V  .4V  57V
#> 48   8V  .5V  58V
#> 49   7V  .4V  52V
#> 50   8V  .4V  49V

Upvotes: 1

see24
see24

Reputation: 1210

You just need to change your pattern to "N|M|P" :

dat <- structure(list(var1 = c("9V", "6V", "9V", "12V", "14V", "13V", 
                        "11V", "11V", "12V", "11V", "13V", "11V", "11V", "9V", "8V", 
                        "4V", "8V", "10V", "9V", "10V", "10V", "11V", "9V", "10V", "7N", 
                        "14N", "7V", "5V", "7V", "9V", "8V", "8V", "5V", "4V", "4V", 
                        "5V", "7V", "5V", "6V", "8V", "9V", "6V", "6V", "7V", "8V", "7V", 
                        "8V", "8V", "7V", "8V"), var2 = c(".6V", ".3V", ".7V", "1.0V", 
                                                          "1.2V", ".8V", ".6V", ".7V", ".8V", ".7V", ".8V", ".7V", ".9V", 
                                                          ".8V", ".7V", ".4V", ".4V", ".8V", ".8V", ".7V", ".6V", ".8V", 
                                                          ".7V", ".7V", ".6N", "1.1N", ".4V", ".3V", ".4V", ".6V", ".5V", 
                                                          ".6V", ".4V", ".3V", ".2V", ".3V", ".4V", ".3V", ".3V", ".5V", 
                                                          ".6V", ".4V", ".4V", ".4V", ".5V", ".4V", ".4V", ".5V", ".4V", 
                                                          ".4V"), var3 = c("77V", "15V", "114V", "199V", "245V", "158V", 
                                                                           "136V", "132V", "171V", "155V", "166V", "138V", "173V", "143V", 
                                                                           "105V", "21V", "26V", "154V", "130V", "113V", "102V", "135V", 
                                                                           "120V", "126V", "124N", "210N", "35V", "9V", "48V", "91V", "81V", 
                                                                           "80V", "14V", "11V", "7V", "13V", "34V", "18V", "15V", "58V", 
                                                                           "76V", "29V", "30V", "31V", "32V", "34V", "57V", "58V", "52V", 
                                                                           "49V")), row.names = c(NA, 50L), class = "data.frame")


library(stringr)
library(dplyr)
dat %>% mutate(var3 = str_replace_all(var3, c("N|M|P"), replacement = NA_character_))

Upvotes: 1

r2evans
r2evans

Reputation: 160437

Here's one solution:

x[] <- lapply(x, function(s) ifelse(grepl("N$", s), NA_character_, s))
x
#    var1 var2 var3
# 1    9V  .6V  77V
# 2    6V  .3V  15V
# 3    9V  .7V 114V
# 4   12V 1.0V 199V
# 5   14V 1.2V 245V
# 6   13V  .8V 158V
# 7   11V  .6V 136V
# 8   11V  .7V 132V
# 9   12V  .8V 171V
# 10  11V  .7V 155V
# 11  13V  .8V 166V
# 12  11V  .7V 138V
# 13  11V  .9V 173V
# 14   9V  .8V 143V
# 15   8V  .7V 105V
# 16   4V  .4V  21V
# 17   8V  .4V  26V
# 18  10V  .8V 154V
# 19   9V  .8V 130V
# 20  10V  .7V 113V
# 21  10V  .6V 102V
# 22  11V  .8V 135V
# 23   9V  .7V 120V
# 24  10V  .7V 126V
# 25 <NA> <NA> <NA>
# 26 <NA> <NA> <NA>
# 27   7V  .4V  35V
# 28   5V  .3V   9V
# 29   7V  .4V  48V
# 30   9V  .6V  91V
# 31   8V  .5V  81V
# 32   8V  .6V  80V
# 33   5V  .4V  14V
# 34   4V  .3V  11V
# 35   4V  .2V   7V
# 36   5V  .3V  13V
# 37   7V  .4V  34V
# 38   5V  .3V  18V
# 39   6V  .3V  15V
# 40   8V  .5V  58V
# 41   9V  .6V  76V
# 42   6V  .4V  29V
# 43   6V  .4V  30V
# 44   7V  .4V  31V
# 45   8V  .5V  32V
# 46   7V  .4V  34V
# 47   8V  .4V  57V
# 48   8V  .5V  58V
# 49   7V  .4V  52V
# 50   8V  .4V  49V

If your data has columns where you do not want to do this replacement, just use a subset:

x[2:3] <- lapply(x[2:3], ...)

Variant:

library(dplyr)
x %>%
  mutate_at(vars(var1, var2, var3), ~ if_else(grepl("N$", .), NA_character_, .))
# or, if all columns
x %>%
  mutate_all(~ if_else(grepl("N$", .), NA_character_, .))

The use of NA_character_ is two-fold:

  1. In the base R version, it is just being declarative, saying that I intend for the result to always be character;
  2. In the dplyr version, its between function requires that the class of both the "yes" and "no" arguments be the same, and class(NA) is not class("A").

Upvotes: 2

Related Questions