Bashscripts.org

I've got a script that will get all the files modified today and I want to copy them to my webserver which has the same directory structure. The only problem I'm having it cp isn't copying it to the directory, its just copying to the root and if I use -r it copies the entire directory. Here's my script so far

find DIR1/* -mtime 0 -name '*' -type f -exec cp -v {} DIR2 \;

any help would be great

I'd do a simple bash for loop to solve this one:



Pay particular attention to the difference between a backtick (` unshifted tilde ~ key) before "find" and after "f"and an apostrophe (' unshifted quote " key) around "*" - it's very important!

that didn't work either. It didn't copy the files into their respective directory in the new folder

uh..... not to be a killjoy, but isn't that what rsync does ?????

Sorry, my command was a bit wrong, but the concept works. This loop actually does what you want it to do:


It'll echo the file being copied to the screen -- remove the first echo to get rid of that.

This script assumes your base directory (DIR1) is different than your target directory DIR2, but the structure of each are identical.

crouse - can rsync just sync after a certain time? I want to be able to just transfer stuff changed today

BrionS,

I'm pretty new to bash scripting, could you explain this line of code:

cp $file `echo $file | sed 's/DIR1/DIR2/'`

Thanks
JGA

Sure thing!

That's a compound line - it's running two commands at once through use of the backticks "`".

Like order of operations - the script will execute what's in the backticks first, and then send its output to the containing statement.

So first it executes

which is a non-elegant way to substitute all the occurances of "DIR1" in the $file (path) with "DIR2". In this case, DIR1 and DIR2 would be replaced with your source and target root folders.

The output of that statement is the modified file path to point to DIR2/path instead of DIR1/path, so the subsequent statement becomes:


The end result is that a file from the source directory that was found by your 'find' command will be copied to the target directory in the same location. Here's an example of a "real" file as it's transformed by this script:


The output of this program will be:

indicating that it copied that file (though it doesn't indicate where to) -- this behavior can easily be modified.

The end result is now you have a copy of the /source/my/dir/testfile in /dest/my/dir/.

So, I'll modify this script to be a little more verbose and a little more parameterized...



Now you can call the program thusly (saved as bsync.sh) using the previous example:


sed is a Stream EDitor that permits you to, among other things, substitute text in a stream with other text -- in this case, the $SRC with the $DEST directory names in the source file path.

Does that help?

thanks BrionS, that explanation helped a lot.

I found a couple of issues. For some reason the $dfile ends up the same as $file so obviously, it doesn't get copied. The other issue I found is if the file name has a space in it. I'll probably just rename those files as file names should probably not have a space in them.

JGA

I found out the sed problem has to do if I pass a / in the directory

so if I do ~/copytest, sed returns an error

I apologize. I wrote a script without testing it.

You're right - it's got all sorts of problems which frankly I don't quite understand. However, I'm pretty sure if you hardcode the source and dest directories you can get it to work (instead of using variables).

Also in sed, to use a '/' character, you need to escape it with '\' such as:


Another thing wrong with my second script is the SRC and DEST need to remove the spaces


As I get time I'll debug the rest - maybe Crouse or jbsnake will pop by and correct my syntax.

again...... i REALLY think you need to "man rysnc" Here is one of the scripts i use......



Daily, this backs up ONLY the CHANGED files from certain directories.... notice the --exclude stuff.
On initial run... it backs up EVERYTHING ....then it only "syncs" the new data from there on out..... very cool, and when used in conjunction with cron, makes for a hands free painless backup setup.

Hope that is of some help

I guess was just hoping for some pointers as to why my script didn't work - rsync aside. I'm not completely familiar with sed yet and from a theoretical standpoint I don't see why this script doesn't at least copy the files it finds.

Consider it practice in script writing - just like writing a compiler to learn how one works, it's not like the world needs another compiler.

the reason I didn't want to use rsync is somebody in a previous post said that rsync doesn't work well syncing two windows boxes (can somebody confirm this?). Another reason is that this isn't a backup script, this is an upload script to my webserver. This lets me quickly get the files I've changed today and upload them.

Thank you to both crouse and BrionS for your scripts.

JGA