[help] How to get current pipeline overall seconds used?

[CorradoLanera]

Help

• I understand and agree to https://books.ropensci.org/targets/help.html.

Description

This is to ask what the suggested way is to calculate the overall amount of time the current (successful) pipeline spends on execution.

My current idea is the following:

tar_meta(fields = c(seconds, warnings, error)) |>
  dplyr::filter(is.na(warnings), is.na(error), !is.na(seconds)) |>
  dplyr::pull(seconds) |>
  sum()

But I'd like to know if it will count targets that are no longer in the pipeline but still stored on meta (i.e., the ones that would be removed by tar_prune()).
Or there could be a more straightforward way to obtain that with a command I did not know/remember. The main point is to keep the "original" time skipped targets spent to be evaluated, too.

Thank you!
Corrado.

UPDATE after answer

So the suggested solution is:

tar_meta(
    tar_manifest()[["name"]],
    fields = c(seconds, warnings, error)
  ) |>
  dplyr::filter(is.na(warnings), is.na(error), !is.na(seconds)) |>
  dplyr::pull(seconds) |>
  sum()

Thank you!

[wlandau-lilly]

This is a bit of a subtly tricky questions because of how parallel computing dynamically schedules the targets as parallel workers become available, but what you propose should be fine if you restrict the tar_meta() output to the targets listed by tar_manifest()$name.