Sort files by filename

Les hlhowell at pacbell.net
Tue Jul 31 19:00:25 UTC 2007


On Tue, 2007-07-31 at 14:50 -0400, Mark Haney wrote:
> Rodolfo Alcazar Portillo wrote:
> > On 7/31/07, Mark Haney <mhaney at ercbroadband.org> wrote:
> >> Rodolfo Alcazar Portillo wrote:
> >>> Am Dienstag, den 31.07.2007, 11:00 -0400 schrieb Miner, Jonathan W (CSC)
> > ...
> >>> [rodolfoap] /home/rodolfoap > for n in $(seq 10 40); do touch XXXX20070515_112011_942_${n}.bz2; done
> >>> [rodolfoap] /home/rodolfoap > rm XXXX20070515_112011_942_11.bz2
> >>> [rodolfoap] /home/rodolfoap > rm XXXX20070515_112011_942_22.bz2
> >>> [rodolfoap] /home/rodolfoap > rm XXXX20070515_112011_942_33.bz2
> >>> [rodolfoap] /home/rodolfoap > for n in $(seq 10 40); do if [ ! -e *_${n}.bz2 ]; then echo NOT FOUND: $n; fi; done
> >>> NOT FOUND: 11
> >>> NOT FOUND: 22
> >>> NOT FOUND: 33
> >>> [rodolfoap] /home/rodolfoap >
> >>>
> >>> Good luck.
> >> I think this would probably work except I really need to know which ones
> >> are out of sequence as is, changing the filename would eliminate that
> >> capability and the missing files are the ones I need.  See, I'm
> >> receiving the files from a separate server and they are sent to me in
> >> this format.  I need to know which ones I /don't/ receive so I know
> >> which ones I need to have resent.  Make sense?
> > 
> > Yes. But you must have a pattern with which we can compare your received
> > files. Which is the exact pattern? If you have this clear, we can
> > probably solve the problem with bash.
> > 
> > Waiting...
> 
> Okay, here goes, the XXXX part denotes a 4 character symbol for a NWS 
> radar site, the 8 digits after that is the date followed by '_HHMMSS_' 
> of time the data was collected.  The next part I'm not certain, but 
> believe it's the scan number of that particular radar scan, then the 
> last 2 digits are the part number of a complete scan.  In other words, I 
> have a series of weather radar scans that get broken up into 'radials' 
> (the .bz2 files) and numbered.  Our customers get fed this data through 
> our servers, but a couple have reported missing files.  I'm trying to 
> track down the problem, but can't go through 2GB of volume scans every 
> day for possibly one missing file.
> 
> The files are stored in directories in this format: XXXX/date/ so I have 
> multiple volume scans in each day's directory.
> 
> Does this help?
> 
> 
> 
> -- 
> Recedite, plebes! Gero rem imperialem!
> 
> 
Hi, Mark, 
	It looks like this is just the sort of situation that rsync was meant
to handle.  You should be able to fire it off with a crontab entry and
have it run periodically.  The rsync command should ensure that all
directories and subdirectories are in sync on the two systems.

Have you looked up its capabilities and thought of how this meshes with
your needs?

Regards,
Les H




More information about the fedora-list mailing list