Search Results

Try:
(./child.ksh >/dev/null) &

If its invoking multiple children then its own PID should be the common PPID of all children...
e.g. here I know I opened all my windows from one, I look for the first and ps grep its PID:
...

pattern_str="1,jobname1"
awk -F, -v td=$temp_date -v pattern_str="${pattern_str}" '$0 ~ pattern_str {$3="completed";$5=td} 1' OFS="," file

You need to use awk variable like this:


awk -F, -v td=$temp_date '/1,jobname1/ {$3="completed";$5=td} 1' OFS="," file

awk -F, 'BEGIN{"date +\"%Y%m%d%H%M%S\""|getline d} /1,jobname1/{$3="completed";$5=d} 1' filename

awk -F, '/1,jobname1,in_progress/ {$3="completed"} 1' filename

You did remove this OFS=","

Try (untested): awk -F, '{$2=$2 FS "status,starttime"} 1' OFS="," file

expr \( $total_size + $file_size \) / 1024
Parentheses are special for the shell. So, you need to escape them.
Also note the whitespace introduced between the parenthesis (both opening and closing)...

With a recent Korn shell (newer than 1988):
total_size=$(( (total_size + file_size) / 1024 ))if you want integer arithmetic or:
total_size=$(( (total_size + file_size) / 1024. ))if you want...

use double quotes


echo "$string"

In order to preserve the tabs, you must double-quote $line.

Simplest solution is to do the splitting with the read statement:
IFS=$'\t' read name age address

If your shell doesn't support that...

Windows uses \r\n as \ line break whereas Unix uses \n. You can pre-filter file like so:

sed 's/$/\r/' <samplefile.txt >sampleflie.win.txt

If your sed doesn't recognise carriage return \r use a tool that does

perl -pi.bak -e 's/$/\r/' path/to/file

It means a pattern which has the value of "par1" followed by 0 or more spaces and 0 or more characters which are not space...

If I understand your requirement properly, try

sed "s|$par1[ ]*[^ ]*|$parval|" sample_file


.* expression will always match the longest possible pattern...need to restrict the matched pattern.

AFAICT, s and b do not need extra arguments, do they?
I would replace this command:
while getopts j:s:b:e: opt; do to while getopts j:sbe: opt; do

Also, try this, for debugging purpose:
s)...

Hi
You can get this in perl:


$ echo "abc efg" | perl -ne 'print s/abc/ABC/; '
1


Substitution in perl returns the number of substitutions made. If none, it will not throw any result.
...

sed '6,$s/<pattern>/<replace_string>/g' inputfile

The shell will not look at your date command which you've protected using single-quotes...

Try this

sed "s/[0-9]\{4\}-[0-9]\{2\}-[0-9]\{2\}$/`date +"%Y-%m-%d"`/g" e123

Check if your input file is in DOS format, it should be in Unix format.

sed 's/[0-9]\{4\}-[0-9]\{2\}-[0-9]\{2\}$/'`date +"%Y-%m-%d"`'/g' sample.txt