CODEWITHSUNDEEP
rvariablesescaping
BioProgram
BioProgram

Reputation: 704

Assign a variable in R using another variable

I have to run 10's of different permutations with same structure but different base names for the output. to avoid having to keep replacing the whole character names within each formula, I was hoping to great a variable then use paste function to assign the variable to the name of the output..

Example:

var<-"Patient1"
(paste0("cells_", var, sep="") <- WhichCells(object=test, expression = test > 0, idents=c("patient1","patient2"))

The expected output would be a variable called "cells_Patient1" Then for subsequent runs, I would just copy and paste these 2 lines and change var <-"Patient1" to var <-"Patient2" [please note that I am oversimplifying the above step of WhichCells as it entails ~10 steps and would rather not have to replace "Patient1" by "Patient2" using Search and Replaced

Unfortunately, I am unable to crate the variable "cells_Patient1" using the above command. I am getting the following error:

Error in variable(paste0("cells_", var, sep = "")) <- WhichCells(object = test, : target of assignment expands to non-language object

Browsing stackoverflow, I couldn't find a solution. My understanding of the error is that R can't assign an object to a variable that is not a constant. Is there a way to bypass this?

Upvotes: 0

Views: 2357

Answers (1)

G. Grothendieck
G. Grothendieck

Reputation: 269654

1) Use assign like this:

var <- "Patient1"
assign(paste0("cells_", var), 3)
cells_Patient1
## [1] 3

2) environment This also works.

e <- .GlobalEnv
e[[ paste0("cells_", var) ]] <- 3
cells_Patient1

3) list or it might be better to make these variables into a list:

cells <- list()
cells[[ var ]] <- 3
cells[[ "Patient1" ]]
## [1] 3

Then we could easily iterate over all such variables. Replace sqrt with any suitable function.

lapply(cells, sqrt)
## $Patient1
## [1] 1.732051

Upvotes: 1

Related Questions