CODEWITHSUNDEEP
r
tzema
tzema

Reputation: 461

R- Check if string contains only certain patterns

I am trying to set up a table where the users can insert formulas in the form of a string.

They are allowed to use only operators and given par names (eg, par.names<- c("P1", "P2", "P3") )

I need to test if the formula is correct, by checking whether a user has used any other patterns.

For example P1+P2*P3 should be acceptable, while P1+P5*P$4 should return an error.

What would be the suggested way to do this?

Thank you!

Upvotes: 3

Views: 1801

Answers (3)

MrFlick
MrFlick

Reputation: 206167

Here's one way, first, you define your variable names and then give a regular expression that would match your operators, then we can make a validater function

par.names<- c("P1", "P2", "P3")
operators <- "[*+]"

get_checker <- function(par.names, operatorsRE) {
  function(x) {
    nonops <- strsplit(x, operatorsRE)  
    sapply(nonops, function(x) all(x %in% par.names))
  }
}

checker <- get_checker(par.names, operators)

We just use an enclosure to keep everything in a tidy function. Then we can use that to validate input

input <- c("P1+P2*P3", "P1+P5*P$4")
checker(input)
# [1]  TRUE FALSE

Here's a version that will allow you to pass an array of operators instead.

par.names<- c("P1", "P2", "P3")
operators <- c("^", "*","+")

get_checker <- function(par.names, operators) {
  operatorsRE <- paste0("(", paste(Hmisc::escapeRegex(operators), collapse="|"),")")
  function(x) {
    nonops <- strsplit(x, operatorsRE)  
    sapply(nonops, function(x) all(x %in% par.names))
  }
}

That way you can avoid regular expression quirks but this does use the Hmisc package for a helper function.

Upvotes: 1

dvd280
dvd280

Reputation: 962

create a vector which holds strings of valid characters:

Allowed = c("+","-","*", "/","P","n","X")

write a function to check if any of the characters in the command are not allowed:

Vcheck = function(string){
  temp = strsplit(string, "")
  (((temp %in% Allowed) %>% sum) == length(temp))  %>% return 
}

This should return 1 if it contains allowed character, and 0 if it contains a forbidden one. I trust you know how to write a function to return an error.

Upvotes: 2

ThomasIsCoding
ThomasIsCoding

Reputation: 101064

Maybe you can try the code below

all(unlist(strsplit(s,"[^[:alnum:]]")) %in% par.names)

Example

> s1 <- "P1+P2*P3"

> s2 <- "P1+P5*P$4"

> all(unlist(strsplit(s1,"[^[:alnum:]]")) %in% par.names)
[1] TRUE

> all(unlist(strsplit(s2,"[^[:alnum:]]")) %in% par.names)
[1] FALSE

Upvotes: 2

Related Questions