Hey all,

I need to do a search and replace on a bunch of files located in different folders. So far I've come up with this solution:

But I'm wondering if there's a better way?

Regards,
Thomas

Looks ok to me. But the only real way to be confident it works is to test it. You could use the File::Find module in Perl (I'm sure there's more than one way to do it with Perl). Depends on whether you're only doing it this once, or if you want to write a generally applicable, robust and reusable script.

What are the brackets {} for? I don't think you need it for sed. Some obscure 'find' syntax? I hate the way you have to add the escaped ';' for an -exec.

does it do what you want? if you have many files to search and replace, you can use xargs instead of -exec in your find command.

"{}" signifies "arguments" passed to -exec command.

2 Perl *so called help* (my cheat sheet(s) if you please) perl scripts. 1 is named clplh (command line perl help) the other is named findh

Since these are made for me to run them from the commandline (and they merely print to STDOUT) it *is* possible that some item(s) are escaped in these scripts so that when they are run that they print correctly (in konsole, I just copy/paste so as to end up with the specific command that I want.

Scripts (not konsole or console output) follow next:

al@AB60R:~/bin$ pwd
/home/al/bin
al@AB60R:~/bin$ cat clplh
#!/usr/bin/perl

print <<STUFF;
# tr/\xA0/ /
perl -pi.bak -e 'tr/\\xA0/ /' filenames*

# all files in a folder
find YourDirectory -type f -print0 | xargs -0 sed -i 's/xxxx/yyyyy/g'

find . | xargs perl -p -i.bak -e 's/oldstring/newstring/g'

perl -e 's/string/stringier/gi' -p -i.bak *.html

via commandline, runs that command, replaces string with
stringier using g global and i insensitive options.
Operates on all .html files in a folder (CWD) also makes
a backup of each file prior to the operation.

STUFF
al@AB60R:~/bin$
al@AB60R:~/bin$
al@AB60R:~/bin$ cat findh
#!/usr/bin/perl
# findh prints example find useage
print "\nfilename accepts wildcards || Or try: locate filename\ncmd | top dir (recurses down) | opt -n denotes-> | name_of_file2find\n\n";
print "find /home -name filename\n--\n";

print <<STUFF;
find / -perm +04000 -exec ls -ld {} \\; 2> /dev/null > /tmp/SUID.files
find / -perm +02000 -exec ls -ld {} \\; 2> /dev/null > /tmp/SGID.files

# files larger than 2 megs. size.txt ordered list
find /home/al/ -xdev -size +2000k -print | xargs ls -ldS > ~/size.txt

# caution searches from / (all of al files on entire disk)
find / -xdev -user al -print | xargs ls -ldS > /home/al/alfil.txt

# change group
find /bin /sbin /usr/bin /usr/sbin -type f -group bin | xargs chgrp root

# all files under yum_repo that are less than 26 days old
find /home/al/yum_repo -mtime -26

# quote when pattern used in name portion
find /mnt/usbhd/rhel/yum_repo -name '*'

find /usr -name foobar.txt

find / -name foobar.txt # srch from root of HD

find / -iname '*.MP3' # case insensitive

find -amin -60 # files accessed less than 60 minutes ago

find / -xdev -ctime -1 # finds new(ly) (software installed) files

perl -MFile::Find -le 'find sub{print if /\\.txt\$/}, "."'

my \@files = `find $dir -name '*.txt'`;

# files only also in cur dir only
find . -name '*' -maxdepth 1 -type f > ~/myfiles.txt
find . -name '*.htm' -maxdepth 1 -type f > ~/myfiles.txt

# cp files to /tmp/slackupdate (ok prompts, exec just does it)
find . -type f -name "*.txt" -maxdepth 1 -ok cp {} /tmp/slackupdate \\;
find . -type f -name "*.txt" -maxdepth 1 -exec cp {} /tmp/slackupdate \\;

locate command. To upgrade the database for the locate command do this while in the superuser mode: slocate -u. Then wait a while while it upgrades. To use the command just type locate mozilla (for example) and it will post all of the mozilla references in your computer.

slocate -u
STUFF
al@AB60R:~/bin$

--
Alan.