CODEWITHSUNDEEP
rfunctiondata.table
Jiexing Wu
Jiexing Wu

Reputation: 313

R data.table function doesn't recognize an already-specified argument

I encounter a weird problem that a data.table function doesn't recognize a well-defined argument if the function is used in another function.

Here is a simple example: I get an error when the first function testFun1,

Error in fun(value) : could not find function "fun"

However, it is clear that there is default value of fun. There is no issue using reshape2::dcast, See testFun2. 

testFun1 <- function(data, formula, fun = sum, value.var = "value") {
  data.table::dcast(data = data, formula = formula, fun.aggregate = fun,
                    value.var = "value")
}

testFun2 <- function(data, formula, fun = sum, value.var = "value") {
  reshape2::dcast(data = data, formula = formula, fun.aggregate = fun,
                    value.var = "value")
}

d <- data.table(x = c("a", "b"), y = c("c", "d"), value = 1)
testFun1(d, x ~ y)
# Error in fun(value) : could not find function "fun"
testFun2(d, x ~ y)

Upvotes: 13

Views: 317

Answers (2)

jangorecki
jangorecki

Reputation: 16697

This issue has been already resolved by recent improvements to dcast by made Arun. They will be soon available on CRAN as 1.12.2 version.

install.packages("data.table", repos="https://Rdatatable.gitlab.io/data.table")
library(data.table)
testFun1 <- function(data, formula, fun = sum, value.var = "value") {
  data.table::dcast(data = data, formula = formula, fun.aggregate = fun,
                    value.var = "value")
}

d <- data.table(x = c("a", "b"), y = c("c", "d"), value = 1)
testFun1(d, x ~ y)

testFun2 <- function(data, formula, fun = sum, value.var = "value") {
  reshape2::dcast(data = data, formula = formula, fun.aggregate = fun,
                  value.var = "value")
}

d <- data.table(x = c("a", "b"), y = c("c", "d"), value = 1)
all.equal(testFun1(d, x ~ y), as.data.table(testFun2(d, x ~ y)), check.attributes=FALSE)
#[1] TRUE

Upvotes: 1

jslice
jslice

Reputation: 21

This seems to be a bug that had existed in a previous version of data.table, was fixed, and has popped up again.

The solution was to change the parameter name in your wrapper function from fun to fun.aggregate so that it matches the name of the data.table::dcast parameter.

Example:

testFun1 <- function(data, formula, fun.aggregate = sum, value.var = "value") {
  data.table::dcast(data = data, formula = formula, 
                    fun.aggregate = fun.aggregate, value.var = "value")
}

Upvotes: 0

Related Questions