Search Results

hi there! you could use find command to find those old mails then remove them..

If you can use cp command first with the "--parents" parameter then you can achieve with ease what you what, then try the find command again then delete them.

find /home/darshak -name "*" -mtime...

If file is consistent, this might work

sed 's/system/HI/2' file

check out diff, comm and even grep utilities..

Did you try changing your prog action into this?

$1 ~ var {print $0}

Your pattern should be ^012$ to get what you want, and your program action should be $1 ~ var {print $0}

What backup scheme are you using? Is it within the LAN or an offsite backup? If I were you, I would discard my current backup process and go with rysnc. It will automate the process of deleting files...

Hi! prior to any deletion, you can try getting all the list of files and redirect it to a text file, then after the deletion or at any given time you might want to get all the list of files again and...

Can you provide your expected output from your given example?

not all the time! if the 1st column has the 789 then your logic will go wrong.. :)

::EDIT:: Haven't read your disclaimer at first.. :)

shorter version of the above suggestion..

awk '/789/ {print $4}' file|wc -l

or simply

sed -n '1,10p' INPUT > OUTPUT && sed -n '30,40p' INPUT >> OUTPUT

What I have in mind is a nasty solution..

1. extract the 1st columns from both files using cut or awk the redirect the output to a temp file.
2. use the suggested grep command above.

...

You might want to post here your sample input files and your expected output file.

A simple

grep -vf fileA fileB

might do the job. You don't need the while statement I guess.

man sed

You might want to try lwp-request. man lwp-request

try this instead

grep -o apple tomo |wc -l

Hi! If grep can be use then try this:

grep apple tomo|wc -w


EDIT: my bad! ignore this suggestion as I tested this with sample file and gives erroneous output.

grep -f set1 set2

I didnt mean the instance of your db in the box. What i meant is the occurences of ORA- in your logfile. In your modified script, as I said it will count all the occurences of ORA- including your...

this will give you erroneous output. you will be counting all instances of ORA- including ORA-00600. try appending a line with ORA-00600 in your logfile and run this command.

The above suggested command will search for the string ORA- excluding instances of "ORA-00600" and will count the total matched lines. Or you could ommit the "wc -l" command and used the "-c" switch...

A simple solution would be to pipe your first grep results to another grep command like this..

grep ORA- alert_pindb.log | grep -v "ORA-00600" |wc -l


Hope that helps.

that's because of bad command substitution plus your expression shouldn't have any spaces on both sides of the operator...

iCOUNT = WC -l COUNTFILE.log

should be

iCOUNT=`WC -l...