Skip to content
Foldr MaSH docs v11.2.0.0

Tidy a folder every night

What it does. Every night at 02:00 the script looks at an Inbox folder, takes the files that haven’t changed for 30 days, and moves them into Archive/<year>. It moves at most 200 files a run, so a big backlog clears over a few nights instead of timing out.

What you need.

  • The MaSH feature.
  • A schedule on Settings > MaSH > Schedule:
    1. Choose Add schedule, then pick the script.
    2. Set Frequency to Daily, Hour to 02:00, and your Timezone.
    3. Under Files to run on, choose Folder listing, then the share and the path (Inbox).
    4. Save.
  • Script settings: ARCHIVE (the archive folder, default Archive) and DAYS (default 30).

A schedule hands the script its files in mash.event.files, once per run. See Schedules.

The decision, on its own

Section titled “The decision, on its own”

The only real logic is “is this old enough?” and “where does it go?”. Both are easy to get wrong with dates, so test them with fixed dates first:

block isOldEnough modified days today
    set cutoff to today.copy().subDays(days)
    return modified.lt(cutoff)
end


block archiveFor root modified
    return "{{root}}/" ~ modified.format("Y")
end


set today to mash.date.parse("2026-11-03")


each [["2026-09-01", "a.pdf"], ["2026-10-25", "b.pdf"], ["2025-12-30", "c.pdf"]] as pair
    set modified to mash.date.parse(pair[0])
    if isOldEnough(modified, 30, today)
        printline "{{pair[1]}}: move to " ~ archiveFor("Archive", modified)
    else
        printline "{{pair[1]}}: keep"
    end
end
a.pdf: move to Archive/2026
b.pdf: keep
c.pdf: move to Archive/2025

copy() matters. Dates in MaSH are mutable, so today.subDays(30) would change today itself. Work on a copy.

The script

Section titled “The script”
set errorblock to "onError"
block onError
    printline "Error on line {{mash.error.line}}: {{mash.error.message}}"
end


block isOldEnough modified days today
    set cutoff to today.copy().subDays(days)
    return modified.lt(cutoff)
end


set archiveRoot to mash.config.get("ARCHIVE") ?? "Archive"
set days to mash.config.get("DAYS") ?? 30
set maxPerRun to 200


set files to mash.event.files
if not files or count(files) is 0
    printline "Nothing in the folder."
    return
end


set today to mash.date.now()
set moved to 0
set kept to 0


each files as doc
    if moved is more than or equal to maxPerRun
        break
    end


    if doc.is_dir
        next
    end


    if not isOldEnough(doc.modified, days, today)
        set kept to kept + 1
        next
    end


    set year to doc.modified.format("Y")
    set yearFolder to mash.files.get(doc.share, archiveRoot).make(year, true)


    # Never overwrite a file already in the archive.
    if mash.files.exists(doc.share, yearFolder.path ~ "/" ~ doc.name)
        printline "{{doc.name}} is already in {{yearFolder.path}}, leaving it."
        next
    end


    doc.move(yearFolder)
    set moved to moved + 1
end


printline "Moved {{moved}}, kept {{kept}}."


if mash.event.files_truncated
    printline "More files matched than were handed over. The next run will carry on."
end


if moved is more than or equal to maxPerRun
    printline "Hit the limit of {{maxPerRun}} for this run. The rest wait for the next one."
end

How it works

Section titled “How it works”
  1. The batch cap (maxPerRun) keeps the run inside its time limit. A queued run has 60 seconds. Moving a file takes a moment, and a large backlog would hit the limit and stop in the middle. See Limits.
  2. Moved files leave the folder. Tomorrow’s listing no longer contains them, so the next run carries on from where this one stopped. That’s the whole trick for big jobs: take a bounded batch, and let the schedule start the next.
  3. files_truncated says when the schedule’s own list was cut at 1,000 files. Both caps lead to the same behaviour, a message and another go tomorrow.
  4. doc.move(yearFolder) takes the folder object, not a path in text. A path would be read on the file’s own share.
  5. next and break skip a file and leave the loop. MaSH calls the first one next, not continue. See Statements.
  6. Idempotent. Running the script twice in one night is harmless: the second run finds less to do.

Variations

Section titled “Variations”
  • Delete instead of archive. Use doc.delete(), but only after you’ve watched the move version behave for a week.
  • Tidy by a field. Use Files matching field criteria in the schedule, for example Status = Closed, and drop the age test.
  • Report what moved. Collect the names in a list and write them to a CSV. See Log each document’s fields to a CSV.
  • Warn before deleting. Run a second, weekly schedule that emails the list of files that will go next week. See Send an email from a template.
  • Stop a second copy starting. Add a script setting MASH_SINGLE_INSTANCE set to true. A new scheduled start is then skipped while the last is still running.

Library pages used: Files, Date, WrappedCarbon.