tables() fails: argument "..." is missing, with no default

[greg-minshall]

the following code generates an error. possibly because the current environment has three dots ("...").

  a <- function(...) data.table::tables()
  a()

the error is

Error in get(x, envir = env) : argument "..." is missing, with no default

# Output of sessionInfo()

R version 4.1.0 (2021-05-18)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: Arch Linux

Matrix products: default
BLAS: /usr/lib/libopenblasp-r0.3.15.so
LAPACK: /usr/lib/liblapack.so.3.10.0

locale:
[1] LC_CTYPE=en_US.UTF-8 LC_NUMERIC=C
[3] LC_TIME=en_GB.UTF-8 LC_COLLATE=en_US.UTF-8
[5] LC_MONETARY=en_GB.UTF-8 LC_MESSAGES=en_US.UTF-8
[7] LC_PAPER=en_GB.UTF-8 LC_NAME=C
[9] LC_ADDRESS=C LC_TELEPHONE=C
[11] LC_MEASUREMENT=en_GB.UTF-8 LC_IDENTIFICATION=C

attached base packages:
[1] stats graphics grDevices utils datasets methods base

loaded via a namespace (and not attached):
[1] compiler_4.1.0 data.table_1.14.3

[MichaelChirico]

Here is a quick fix in case you're stuck:

a <- function(...) data.table::tables(env = .GlobalEnv)
a()
# No objects of class data.table exist in .GlobalEnv 

[MichaelChirico]

BTW, my gut feeling is that env = .GlobalEnv is closer to what you're after anyway. When used in a function like you've provided, by default tables() will only list the tables in that function's environment.

[greg-minshall]

hi, Michael, thanks for the quick response (and already a pull request). in fact, i'm trying to walk back the call stack, finding the data tables in each frame. (i have a process that runs for an hour or so, very little lives in the global environment, and i'm curious about memory usage.) at the end, i do some lame unique( as a hack to get rid of duplicates.

just since i have you, if you've any idea on how to find aliases (to de-duplicate), i'm all ears. fyi, this is the code so far:

  dts <- NULL
  eh <- environment()
  for (i in 1:sys.nframe()) {
    dts <- rbind(dts,
                 tables(env = sys.frame(i), silent = TRUE))
  }
  dts <- data.table(unique(as.data.frame(dts)))

where the "shell game" on the last line is because, amusingly, the data.table that tables() returns seems to not be unique'i'fiable by data.table. :) because, i guess, of the list of column names.

[MichaelChirico]

Some tips:

(1) replace COLS := sapply(COLS, toString) to get unique.data.table working again

(2) if you're looking for true duplicates (sharing the same memory), you might look at the address() of each table you find:

# within the loop
dts[, address := sapply(NAME, function(nm) address(sys.frame(i)[[nm]]))]

ought to work.

(3) you might want idcol = TRUE or idcol = 'frame' to record which frame each table came from

[greg-minshall]

Michael, thanks for the tips. address(), in particular, i'm very happy to know. my code (just because) is now:

  exceptions <- 0
  dts <- rbindlist(
    lapply(
      seq_len(sys.nframe()),
      function(i) {
        ##  deal with a bug in data.table
        ##  https://github.com/Rdatatable/data.table/issues/5197
        dt <- tryCatch(
          dt <- tables(env = sys.frame(i), silent = TRUE),
          error = function(e) {
            exceptions <<- exceptions + 1L
            return(data.table())
          })
        if (nrow(dt)) {
          return(data.table(dt, ID = i))
        } else {
          return(NULL)
        }
      }))

(exceptions is to track how many frames i can't examine; assuming your pull makes it into the development branch (i'm addicted to env = list(...)), i can simplify.)

cheers.

[greg-minshall]

thanks very much -- and for the new dev release with this fix. (happiness is simplified code! :)

cheers.