How to do a for loop with case_when - r

Question

I'm a beginner with R and I'm trying to do a for-loop to recode many variables: when "test" modality is missing, then have "test.v1" modality. It looked very easy to do, but I can't get it:
VEC_1 <- c("test1","test2","test3","test4","test5","test6","test7","test8","test9")
VEC_2 <- c("test1.v1","test2.v1","test3.v1","test4.v1","test5.v1","test6.v1","test7.v1","test8.v1","test9.v1")
for (i in 1:(min(length(VEC_1), length(VEC_2)))){
df2 <- df1 %>%
mutate(
VEC_1[i] = case_when(
is.na(VEC_1[i]) & !is.na(VEC_2[i]) ~ VEC_2[i],
TRUE ~ VEC_1[i])
)
}
I have this error
Unexpected error : '=' in:
" mutate(
VEC_1[i] ="
Does anyone have an idea ?
EDIT: df1 is like :
test1 <- c("A","B","A","A",NA,"B","A",NA,"A")
test1.v1 <- c("B",NA,"B","B","A","B","B",NA,"A")
test2 <- c("B","B","B","B",NA,"C","C","C","C")
test2.v1 <- c("C",NA,"A","A","B","B","C",NA,"C")
test3 <- c("A","B","B","B",NA,"C","C",NA,"C")
test3.v1 <- c("B","A","B",NA,"A","A","A","A",NA)
test4 <- c(NA,"B","B","A",NA,"B","A",NA,"A")
test4.v1 <- c("B","B","B","A","A","B","B","B","B")
df1 <- data.frame(test1,test1.v1,test2,test2.v1,test3,test3.v1,test4,test4.v1)

Answer 1

Based on the example data.frame df1, I'm wondering if you might try putting your data into long form, then grouping by row number and test, then substituting missing values.
library(tidyverse)
df1 %>%
mutate(rn = row_number()) %>%
pivot_longer(cols = -rn, names_to = c("test", "mode"), names_pattern = "test(\\d+)([.v1]*)") %>%
group_by(rn, test) %>%
mutate(value = ifelse(mode == "" & is.na(value), value[mode == ".v1"], value)) %>%
pivot_wider(id_cols = rn, names_from = c(test, mode), values_from = value, names_prefix = "test", names_sep = "")
Output
rn test1 test1.v1 test2 test2.v1 test3 test3.v1 test4 test4.v1
<int> <chr> <chr> <chr> <chr> <chr> <chr> <chr> <chr>
1 1 A B B C A B B B
2 2 B NA B NA B A B B
3 3 A B B A B B B B
4 4 A B B A B NA A A
5 5 A A B B A A A A
6 6 B B C B C A B B
7 7 A B C C C A A B
8 8 NA NA C NA A A B B
9 9 A A C C C NA A B

Answer 2

I'm assuming VEC_1 and VEC_2 are the same length here. you don't need to use a for-loop, mutate affects the whole column of the dataframe, it kinda behaves like a for-loop. I have rewritten your code like this and changed the data for testing purposes:
library(dplyr)
VEC_1 <- c("test1","test2","test3","test4",NA,"test6","test7",NA,"test9")
VEC_2 <- c("test1.v1",NA,"test3.v1","test4.v1","test5.v1","test6.v1","test7.v1",NA,"test9.v1")
df <- data.frame(VEC_1,VEC_2)
df %>% mutate(VEC_1 = if_else(is.na(VEC_1) & !is.na(VEC_2),VEC_2,VEC_1))
which is equal to
df <- data.frame(VEC_1,VEC_2)
for (i in 1:nrow(df)){
if(is.na(df$VEC_1[i]) & !is.na(df$VEC_2[i])){
df$VEC_1[i] = df$VEC_2[i]
}
}
Output:
> df
VEC_1 VEC_2
1 test1 test1.v1
2 test2 <NA>
3 test3 test3.v1
4 test4 test4.v1
5 test5.v1 test5.v1
6 test6 test6.v1
7 test7 test7.v1
8 <NA> <NA>
9 test9 test9.v1
i also changed from case_when to if_else, because you only check one condition.
?if_else
if_else(condition, true, false, missing = NULL)

Answer 3

ifelse/if_else will be a good solution, unless you have multiple conditions, then case_when is your friend. A loop isn't necessary:
library(dplyr)
VEC_1 <- c("test1","test2",NA,NA)
VEC_2 <- c("test1.v1",NA,NA,"test4.v1")
df <- tibble(VEC_1, VEC_2)
df %>%
mutate(VEC_1 = case_when(
is.na(VEC_2) ~ VEC_1,
is.na(VEC_1) ~ VEC_2,
TRUE ~ VEC_1)
)
# A tibble: 4 × 2
VEC_1 VEC_2
<chr> <chr>
1 test1 test1.v1
2 test2 NA
3 NA NA
4 NA test4.v1
# A tibble: 4 × 2
VEC_1 VEC_2
<chr> <chr>
1 test1 test1.v1
2 test2 NA
3 NA NA
4 test4.v1 test4.v1