grepping the smart way

Unfortunately WebLogic prints each log entry in multiple lines. This is a pain when you grep.

gawk will return the whole log entry:

gawk "BEGIN{RS=\"####\"}/BEA-000337/{print \$0}" *.log

When using grep, the option -A4 is handy to print a few extra lines (in the example, 4) after the ling matching the pattern:

grep -A4 BEA-000337 *

Bash shell tricks

Tarring files recursively with filter

to create:
find . -name *.out | xargs tar cvf cohservernodes.tar

to append:
find . -name *.log | xargs tar rvf cohservernodes.tar

to list:
tar tf cohservernodes.tar


Find a text in files, recursively:

find . | xargs grep -s whatever



find and replace recursively

for file in $(grep -il "whatever" *.txt)
do
sed -e "s/Hello/Goodbye/ig" $file > /tmp/tempfile.tmp
mv /tmp/tempfile.tmp $file
done

how about:

perl -p -i -e 's/oldstring/newstring/g' `find ./ -name *.html`

the above command works very well, but if you try

perl -p -i -e 's/oldstring/newstring/g' `find ./ -name *`

you get the error:

Can't do inplace edit: bla is not a regular file, <> line

because it tries to change also the directories

To do a find excluding directories you must do this (add -type f):

perl -p -i -e 's/oldstring/newstring/g' `find ./ -name * -type f`