How to populate a coll with a list of files?


    Apr 01 2006 | 1:34 pm
    I need to populate a coll with a filenames of files contained in a folder.
    For now, I use ubumenu (with dropfile) to get the list of files and then I parse them one by one to get them into coll. I know there are very handy third party objects that do this kind of stuff, but for this I want to use only standard distribution objects.
    What I did works, but from time to time there is some parsing error (like non exiting files appearing in a list), that i didn't manage to nail yet. I will make it work, but i was wondering if there was some other automatic and more reliable method to acomplish this.
    I would be very happy if I could somehow read the content of ubumenu, but that doesn't seem to be possible.
    I suppose this kind of thing could be easily done with Java, but I don't know anything about that stuff.
    Klif

    • Apr 01 2006 | 5:06 pm
      can you post the patch, so the hive mind can take a look and see where the error might be?
      v a d e //
      www.vade.info abstrakt.vade.info
    • Apr 02 2006 | 12:22 pm
      thanx vade.
      My prime concern was if there is some easier way to acomplish this.
      Thinking about the problem I realised I was parsing using metro to drive it. I think i should change this and bang the next read (from ubumenu) when previous file has finished parsing.
      Also, there might be some problems with patch initialisation.
      I'll clean the patch and post the results/problems.
      Again, thanx for your concern.
      Klif
    • Apr 02 2006 | 8:19 pm
      Hello klif, this is how I once did it, using the folder object. this was a while back, and only used on a mac. so lots of disclaimers here. I am not sure if this addresses your "parsing errors" problem. HtH anyway. -jennek
    • Apr 03 2006 | 4:38 pm
      Thanks Jennek, and vade!
      I completely forgot about folder object. I was using it before, but since ubumenu appeared, it seems that folder object disappeared from my fragile memory ;)
      As for parsing errors that i was mentioning before, I figured what was wrong (I still don't know why tough).
      I was using a feedback loop to feed the counter at the "top" of the patch. Counter would trigger ubumenu/coll filled with filenames. Then I would do some processing on the filename (extracting date and time values from filenames *** see below), saving results in second coll.
      Obviously, counter was banged by last operation in the chain (in this case, it was parsed data going to second coll). For some reason, the loop would stop ocasionally, and I needed to add considerable delay (and different amounts => delay 3 on dual 2Gig G5, delay 10 on dual 1Gig G4, delay 15 on Titanium 400mhz) to get all the values.
      After some messing with the patch, I ended up using deferlow followed by delay in feedback loop. Now the patch works sucessfully on all machines I tried it on.
      I seem to remember that I used to have similar problems before, when I would be using long lists stored in colls.
      It would probably be easier if I posted the simplified patch in the first place, but I don't have license for max on my old laptop that I use to check mail, so I can't easily prepare and send patches.
      Klif
      *** This is the part of the project I don't know much about. Some kind of video server, records video into mpeg files, and names them by date and time (like VIDEO_20061015_234512.mpg YYYYMMDD_HHMMSS format). My friend wants to read specific files/time and do some processing on it.
      I need data in format YMMDDHHMMSS - in that case, I can directly compare values using ">" and " specific "moment in time".