Use result of subquery in conditional mutate - r

Question

I want to compute a new row based on the result of a subquery in the same dataframe. Minimal (non) working example:
library(plyr)
library(dplyr)
df <- data.frame(
VAR1 = c("A", "A", "B", "C"),
VAR2 = c("F", "G", "E", "D"),
VAR3 = c("G", "F", "X", "D")
) %>% as_tibble
subquery <- function(v1, v2) {
dplyr::filter(df, as.character(v1) == VAR1, as.character(v2) == VAR2)
}
TEST <-
df %>%
mutate(X = case_when(
plyr::empty(subquery(VAR1, VAR3)) ~ "EMPTY",
TRUE ~ "NON EMPTY"
))
The result dataframe TEST should be
VAR1 VAR2 VAR3 X
<fctr> <fctr> <fctr> <chr>
A F G NON EMPTY
A G F NON EMPTY
B E X EMPTY
C D D NON EMPTY
but is
VAR1 VAR2 VAR3 X
<fctr> <fctr> <fctr> <chr>
A F G NON EMPTY
A G F NON EMPTY
B E X NON EMPTY
C D D NON EMPTY
Many thanks in advance!
remark: If I don't coerce the v1 and v2 to character I get the following error:
Error in mutate_impl(.data, dots) :
Evaluation error: Evaluation error: level sets of factors are different..

Answer 1

I would put the empty function within the subquery function in order to return TRUE or FALSE values. Then it would be possible to vectorise it in order to apply it to every row of your dataframe:
library(plyr)
library(dplyr)
df <- data.frame(
VAR1 = c("A", "A", "B", "C"),
VAR2 = c("F", "G", "E", "D"),
VAR3 = c("G", "F", "X", "D")
) %>% as_tibble
subquery <- function(v1, v2) {
empty(filter(df, as.character(v1) == VAR1, as.character(v2) == VAR2))
}
subquery = Vectorize(subquery)
df %>%
mutate(X = case_when(
subquery(VAR1, VAR3) == FALSE ~ "NON EMPTY",
TRUE ~ "EMPTY"
))
# # A tibble: 4 x 4
# VAR1 VAR2 VAR3 X
# <fct> <fct> <fct> <chr>
# 1 A F G NON EMPTY
# 2 A G F NON EMPTY
# 3 B E X EMPTY
# 4 C D D NON EMPTY
Or you can put empty and case_when within the subquery function like this:
subquery <- function(v1, v2) {
res = empty(filter(df, as.character(v1) == VAR1, as.character(v2) == VAR2))
case_when(res == FALSE ~ "NON EMPTY",
TRUE ~ "EMPTY")
}
subquery = Vectorize(subquery)
df %>% mutate(X = subquery(VAR1, VAR3))