Allow for depends files to be a list of character vectors?

[kendonB]

Currently, it looks like the commands function with depends works something like this:

datasets <- commands(
  `filename.rds` = file_maker(target_name))
datasets$depends <- large_vector_of_file_paths %>% paste0(collapse = ", ")

Printing this data.frame as a tibble doesn't look nice. It seems like a more natural way to do this is to allow the large_vector_of_file_paths to be a list containing that vector. As in:

datasets$depends <- list(large_vector_of_file_paths)

Seems tidier? Am I doing this wrong? I want filename.rds to depend on a large number of files.

[wlandau]

Nice idea. To be honest, I have been resisting it because I like the idea of forcing YAML-like lists to be flat, but I agree that your proposal is more natural and tidy for depends. I think it should be possible to accommodate both kinds of input.

In general, I have sometimes been thinking about with how to display workflow plan data frames with unavoidably long entries, such as complicated commands. I am not sure tibble's print method is adequate because it seems to skip cumbersome columns. Would you be interested in a special S3 print method that shows truncated versions of all or most entries?

Also, I am curious as to why your use case requires a large number of files. Are you feeding the output into another program? Personally, I love that remake keeps non-file targets in a storr cache so I do not have to worry about generating files most of the time.

[kendonB]

The large number of files are weather data. I have no control over the format. Though, a large number of small files is better for the provider as it doesn't force people who want a single day of data to download 50 years.

[kendonB]

Also, I don't think it's worth it to make another print method for these. I'm really only going to print these to make sure the generating code worked and seeing a vector of the correct length is good enough for me.