eplusr error on upgrade data.table #4612
Comments
|
This is a regression. Still haven't quite nailed down an MRE but it's do do with using a variable in j from the same environment, which is causing the |
|
I am the author of eplusr package. Maybe I can help to provide a MRE here. But I did not know how to install the version that cause this error. Can I directly install the dev version of data.table via |
|
@hongyuanjia That is correct. The release candidate version in current development is in the main branch, and with that |
|
Thanks @hongyuanjia , yes the latest dev reproduced the error for me. However it stopped reproducing as the |
I can recommend registering a GITHUB_PAT and setting it as an environment variable. |
|
Below is a MRE. Seems like a scoping issue: library(data.table)
dt <- data.table(id = c(1, 1, 2), value = c("a", "b", "c"))
fun <- function (dt, tag = c("A", "B")) {
dt[, var := tag[[.GRP]], by = "id"]
}
fun(dt)
#> Error in eval(call("is.atomic", jsub[[2L]]), envir = x): object 'tag' not foundCreated on 2020-07-20 by the reprex package (v0.3.0) |
|
Interesting. Inside library(data.table)
dt <- data.table(id = c(1, 1, 2), value = c("a", "b", "c"))
fun1 <- function (dt, tag = c("A", "B")) {
# using "[[" does not work
dt[, var := tag[[.GRP]], by = "id"]
}
fun1(dt)[]
#> Error in eval(call("is.atomic", jsub[[2L]]), envir = x): object 'tag' not found
fun2 <- function (dt, tag = c("A", "B")) {
# using "[" works
dt[, var := tag[.GRP], by = "id"]
}
fun2(dt)[]
#> id value var
#> 1: 1 a A
#> 2: 1 b A
#> 3: 2 c BCreated on 2020-07-20 by the reprex package (v0.3.0) |
|
The error comes from: Line 1545 in 1a8a390 where |
|
I can confirm that all errors are caused by this issue. After I temporarily getting workaround on this, all check errors are gone and eplusr can work well with dev version of data.table. |
|
Thanks @hongyuanjia for the MRE. I do think |
|
@MichaelChrico I looked into this - it looks like simply adding |
|
Thanks @ColeMiller1 I had tried that but I only tried it on a non-MRE so it didn't work. Glad it's as easy as that.
|
The text was updated successfully, but these errors were encountered: