CODEWITHSUNDEEP
rlogictidyversetruthtable
knb
knb

Reputation: 9313

3-value truth-table with R functions. conditional?

I want to construct a truthtable with R tidyverse functions mutate(), case_when() and expand_grid().
Here I define a custom operator %->% for the "conditional" logical connective
(->, "if p then q" AKA "p imples q".)
I also try to consider R's three-value logic.

Is the following code correct?
Is my %->% already taken by some other common package?

library(tidyverse)
values <- c(TRUE, FALSE, NA)

`%->%` <- function(p,q) {
  case_when((p & q) ~ TRUE,
            (p & !q) ~ FALSE,
            (p & is.na(q)) ~ TRUE,


            (!p & q) ~ FALSE,   # not sure about this one
            (!p & !q) ~ TRUE,
            (!p & is.na(q)) ~ NA,


            (is.na(p) & q) ~ NA,
            (is.na(p) & !q) ~ TRUE,
            (is.na(p) & is.na(q)) ~ NA,

            #unhandled case, this line should never be reached
            TRUE  ~ NA )}


tt <- expand_grid(p = values, q = values) %>%
  mutate(`p|q` = p | q,
         `p&q` = p & q,
         `p->q` = p %->% q) %>%
  select(p, q, `p->q`)
tt

Result:

R> tt
# A tibble: 9 × 3
  p     q     `p->q`
  <lgl> <lgl> <lgl> 
1 TRUE  TRUE  TRUE  
2 TRUE  FALSE FALSE 
3 TRUE  NA    TRUE  
4 FALSE TRUE  FALSE  # not sure about this one
5 FALSE FALSE TRUE  
6 FALSE NA    NA    
7 NA    TRUE  NA    
8 NA    FALSE TRUE  
9 NA    NA    NA

Is there a built-in R operator (or a go-to package on CRAN ) for the conditional that I don't know about (faster, more rigorously tested, vectorized, generic...)?

(I have intermediate R skills and I am a beginner in logic)

Update 2022 : I have rewritten the function/truthtable according to a Book (J.Celko, Data&Databases, Morgan Kaufman Publ, 1999)

Upvotes: 0

Views: 114

Answers (0)

Related Questions