Hi,

I would like some help on some shell scripting that I am trying to do.

Basically, I would like to write a shell script to ftp a file and then rename

the file to something else.

I already have a shell script to ftp a file if the file name does not change.
However, the file I want to ftp now has a file name change to the last
part of the file name every month.

For example:

newfile.20050801.gz will be newfile.20050901.gz next month to reflect
that the file has been updated and was created in september.

I want to write some lines that will look into the directory of interest and
grab the file name of the newest file there and insert it into a variable.

Then I want to rename the file.

Also, as an aside how do I capture the date a file was last saved in a variable?

How would I do this??

thanks very much!!

You might try using sed and regular expressions.

Not sure what you mean by 'catpure' but something like

$variable > filename

will take the variable and place it in a file, not sure if that is what you wanted.

HTH

I would recommend renaming the file locally. You may overwrite remote files otherwise. If you keep a local copy of everything that exists remotely (or at least a file containing a list of it), you can track what's there without connecting to the remote machine just to list files, and you'll never accidentally rename a file to overwrite an existing one.

Well, not never, but you know... ;)

The 'ls' command can list files in a directory in order of modification time. It will also list mod time with the -l flag. Man ls for more info.

Try
ls -ltra

Play with the below commands in the directory, and see what happens when you change the number 6 to 7 or 8:
ls -laF | awk '{print $6}'
ls -ltra | tail -5 | awk '{print $7}'

[edit:] One more hint:
FILENAME=`ls -ltra | tail -1 | awk '{print $8}'`
will set $FILENAME to name the last modified file. Careful, because it is likely to be the working directory.

Hi,

Thanks for the suggestions.

I tried the command below and it gave me a time of day. I tried to remove the r and it gave me the year of the oldest file.
FILENAME=`ls -ltra | tail -1 | awk '{print $8}'`


I am confused because the first thing I am trying to do is to capture the name of the newest file from the
site I am trying to grab a file of.

On my own directory I am trying to grab the time stamp of a file that I do know the name of.

I tried the other examples but I seem to be getting parts of a date I acutually want the whole date.

thanks very much!!

I would (if possible) use scp if I were you

scp <local-file> <user>@<host>:<remote-file>

If you set-up passowrd less authentication for ssh
scripting it is trivial.

Sorry, I haven't had time to check my mail.
The year of the oldest file. Correct. So the time of day was the last time the NEWEST file in the directory was modified. The different flags on this usage of ls are:
a - all files
l - long listings
t - sort by last modification time
r - reverse order

Piping the list (or anything else) through tail gives the "tail end" of the output. The default is ten lines, and we ask for one line with "-1." Piping THAT through this simple usage of awk (I am not an awk guru) will simply give you a field (in this case the 8th) in the line. Playing around with that number should show you a bit more.

Try eliminating the tail and/or printing more fields. Depending on how many files are in your directory, you should get a long list.
'ls -ltra | awk '{print $5" "$6" "$7}'

Well, depending on how you connect, this may become far more complicated.

To do it locally, start simple:With the awk stuff, you can see exactly when each file was modified. I suppose you wouldn't need awk

By the way, I strongly agree with Tinkster. SSH is a far better alternative to FTP if you don't have a strong reason to NOT use a secure connection and transfer.

Also, depending on your situation, you might do a bit of research on the rsync command. It will synchronize a remote directory with a local one (using your choice as the master) and can use the ssh protocol as well.

Above all, playing with stuff is how you learn. Copy a bunch of your important files into a new experimental directory so you know you have duplicates. Move things. Rename them. List them and sort them. Man tar. The more you ask yourself "I wonder what happens if I...?" the more you'll learn about things that will prove useful sooner or later. If you have everything backed up, you can be adventurous knowing you won't break anything that can't be restored.

I initially assumed you wanted to UPload this file. I realize now that I was probably mistaken about that. I was going to clarify a bit more correct spelling, clumsy phrasing, etc), but I just got a call that my son is sick, so I'm stopping now because posting what's here will help more than no reply. I hope this helps.