有没有办法根据文件的存在而不是特定时间来触发 cron 作业运行？

Question

I am trying to use cron r or task scheduler in R to run a script daily based off a.CSV file that gets updated everyday. The one thing is there is no specific time of the day the CSV file gets updated (let's say on 4/20 it got updated at 3PM but at 4/21 it got updated at 2:30PM and at 4/22 it got updated at 12PM). The main trigger is not time of day but daily existence of file. Is there a way I can run this using either of the R addins? I use a server at work so I am not using windows task scheduler since R is not on my machine.

Answer 1

Instead of running the cron job every day, run it every 5 minutes (or some reasonable interval), and keep track of when it processed the file. For example,

needswork <- function(filename, expr, updated = paste0(filename, ".seen")) {
  if (!file.exists(filename)) return(FALSE)
  if (!file.exists(updated)) return(TRUE)
  return(file.info(updated)$mtime < file.info(filename)$mtime)
}
donework <- function(filename, expr, updated = paste0(filename, ".seen")) {
  writeLines(character(0), updated)
}

if (needswork("/path/to/mainfile.csv")) {
  # process the file here
  # ...
  # update
  donework("/path/to/mainfile.csv")
}

I might extend the needswork a little to add notification problems, such as

needswork <- function(filename, expr, updated = paste0(filename, ".seen")) {
  if (!file.exists(filename)) return(FALSE)
  if (difftime(Sys.time(), file.info(filename)$mtime, units="secs") > 60*60*24) {
    some_notify_function()
    # perhaps something like
    msg <- paste("The file", sQuote(filename), "has not been updated since",
                 file.info(filename$mtime))
    RPushbullet::pbPost("note", title = "No recent updates", body = msg)
  }
  if (!file.exists(updated)) return(TRUE)
  return(file.info(updated)$mtime < file.info(filename)$mtime)
}

Answer 2

Cron is strictly a time based scheduler.

Having said that, there is a work around.

1. Create a script (eg: mycron.py) as follows

import os.path

if os.path.isfile("/tmp/myfile.csv"):
  # File exists
  # Do something
else:
  # File does not exist
  pass

2. Schedule this script (mycron.py) to run at regular intervals

Python script is just an example. Feel free to use your fav scripting language