Optimize shell one liner - shell

Need to optimize the UNIX shell one liner
cat ${TEMPFILE} | cut -d ' ' -f1 | sed '/^$/d'| sed '1,4d'| sed 's/$/|ON_ICE|OFF_ICE/g' > ${MYREPORT}
as this is causing performance issues.

Call sed only once:
cat ${TEMPFILE}|cut -d ' ' -f1|sed '/^$/d;1,4d;s/$/|ON_ICE|OFF_ICE/g'>${MYREPORT}

use awk as follows:
awk '{$0=$1};if (NF>1){++rec}; if(NF > 1 && rec > 4 ){sub(/$/,"|ON_ICE|OFF_ICE")); print};' ${TEMPFILE} > ${MYREPORT}

awk '/^$/ || ++count <= 4 {next} {print $1 "|ON_ICE|OFF_ICE"}' "$TEMPFILE" > "$MYREPORT"

In
cat ${TEMPFILE} | cut -d ' ' -f1 | sed '/^$/d' | sed '1,4d' | sed 's/$/|ON_ICE|OFF_ICE/g' > ${MYREPORT}
clearly you can replace sed '1,4d' with tail +4

Related

Print output of two commands in one line

I'm got this working:
while sleep 5s
do
lscpu | grep 'CPU MHz:' | cut -d ':' -f 2 | awk '{$1=$1};1' && grep 'cpu ' /proc/stat | awk '{usage=($2+$4)*100/($2+$4+$5)} END {print usage "%"}'
done
And it gives me the following output:
1601.058
3.4811%
1452.514
3.48059%
1993.800
3.48006%
2085.585
3.47955%
2757.776
3.47902%
1370.237
3.47851%
1497.903
3.47798%
But I'd really like to get the two values onto a single line. Every time I try to do this I run into a double / single quote variable issue. Granted I pulled some of this awk stuff from online so I'm not really up to speed on that. I just want to print per line, CPU clock and load ever 5 seconds.
Can you help me find a better way to do that?
You may use process substitution to run lscpu and cat /proc/stat and feed to single command. No need to use pipes.
while sleep 5; do
awk '/CPU MHz:/{printf "%s ", $NF} /cpu /{print ($2+$4)*100/($2+$4+$5)"%"}' <(lscpu) /proc/stat
done
If there is only one input command:
date| awk '{print $1}'
Wed
OR
awk '{print $NF}' <(date)
2019
If more then one command: Example , get the year of of the two date command in same line. (not very useful example, only for sake of demo)
awk '{printf "%s ", $1=NF}END{print ""}' <(date) <(date)
2019 2019
pipe the output of the 2 commands into paste
while sleep 5; do
lscpu | awk -F':[[:blank:]]+' '$1 == "CPU MHz" {print $2}'
awk '$1 == "cpu" {printf "%.4f%%\n", ($2+$4)*100/($2+$4+$5)}' /proc/stat
done | paste - -
The 2 columns will be separated by a tab.
Writing this for readability rather than efficiency, you might consider something like:
while sleep 5; do
cpu_pct=$(lscpu | awk -F': +' '/CPU MHz:/ { print $2 }')
usage=$(awk '/cpu / {usage=($2+$4)*100/($2+$4+$5)} END {print usage "%"}' /proc/stat)
printf '%s\n' "$cpu_pct $usage"
done
Command substitutions implicitly trim trailing newlines, so if lscpu | awk has output that ends in a newline, var=$(lscpu | awk) removes it; thereafter, you can use "$var" without that newline showing up.
All you need to do is change the newline on the first line to a different separator. Something like:
lscpu | ... | tr \\n : && grep ...
You can also use echo -n $(command_with_stdout). The -n switch specifies that the new line (\n) will be omitted.
while sleep 5s; do
echo -n $( lscpu | grep 'CPU MHz:' | cut -d ':' -f 2 | awk '{$1=$1};1' )
echo -n ' **** '
echo $( grep 'cpu ' /proc/stat | awk '{usage=($2+$4)*100/($2+$4+$5)} END {print usage "%"}' )
done
Or the same representation in one line:
while sleep 5s; do echo -n $( lscpu | grep 'CPU MHz:' | cut -d ':' -f 2 | awk '{$1=$1};1' ); echo -n ' **** '; echo $( grep 'cpu ' /proc/stat | awk '{usage=($2+$4)*100/($2+$4+$5)} END {print usage "%"}' ); done
EDIT: (remove -n switch from echo according to Charles Duffy's comment)
while sleep 5s; do echo "$( lscpu | grep 'CPU MHz:' | cut -d ':' -f 2 | awk '{$1=$1};1' ) **** $( grep 'cpu ' /proc/stat | awk '{usage=($2+$4)*100/($2+$4+$5)} END {print usage "%"}' )"; done

Count the number of whitespaces in a file

File test
musically us
challenged a goat that day
spartacus was his name
ba ba ba blacksheep
grep -oic "[\s]*" test
grep -oic "[ ]*" test
grep -oic "[\t]*" test
grep -oic "[\n]*" test
All give me 4, when I expect 11
grep --version -> grep (BSD grep) 2.5.1-FreeBSD
Running this on OSX Sierra 10.12
Repeating spaces should not be counted as one space.
If you are open to tricks and alternatives you might like this one:
$ awk '{print --NF}' <(tr -d '\n' <file)
11
Above solution will count "whitespace" between words. As a result for a string of 'fifteen--> <--spaces' awk will measure 1, like grep.
If you need to count actual single spaces you can use this :
$ awk -F"[ ]" '{print --NF}' <<<"fifteen--> <--spaces"
15
$ awk -F"[ ]" '{print --NF}' <<<" 2 4 6 8 10"
10
$ awk -F"[ ]" '{print --NF}' <(tr -d '\n' <file)
11
One step forward, to count single spaces and tabs:
$ awk -F"[ ]|\t" '{print --NF}' <(echo -e " 2 4 6 8 10\t12 14")
13
tr is generally better for this (in most cases):
tr -d -C ' ' <file | wc -c
The grep solution relies on the fact that the output of grep -o is newline-separated — it will fail miserably for example in the following type of circumstance where there might be multiple spaces:
v='fifteen--> <--spaces'
echo "$v" | grep -o -E ' +' | wc -l
echo "$v" | tr -d -C ' ' | wc -c
grep only returns 1, when it should be 15.
EDIT: If you wanted to count multiple characters (eg. TAB and SPACE) you could use:
tr -dC $'[ \t]' <<< $'one \t' | wc -c
Just use awk:
$ awk -v RS=' ' 'END{print NR-1}' file
11
or if you want to handle empty files gracefully:
$ awk -v RS=' ' 'END{print NR - (NR?1:0)}' /dev/null
0
The -c option counts the number of lines that match, not individual matches. Use grep -o and then pipe to wc -l, which will count the number of lines.
grep -o ' ' test | wc -l

using date variable inside sed command

I am storing date inside a variable and using that in the sed as below.
DateTime=`date "+%m/%d/%Y"`
Plc_hldr1=`head -$i place_holder.txt | tail -1 | awk -F ' ' '{ print $1 }'`
Plc_hldr2=`head -$i place_holder.txt | tail -1 | awk -F ' ' '{ print $2 }'`
sed "s/$Plc_hldr1/$DateTime/;s/$Plc_hldr2/$Total/" html_format.htm >> /u/raskar/test/html_final.htm
While running the sed command I am getting the below error.
sed: 0602-404 Function s/%%DDMS1RT%%/01/02/2014/;s/%%DDMS1C%%/1235/ cannot be parsed.
I suppose this is happening as the date contains the following output which includes slashes '/'
01/02/2014
I tried with different quotes around the date. How do I make it run?
Change the separator to something else that won't appear in your patterns, for example:
sed "s?$Plc_hldr1?$DateTime?;s?$Plc_hldr2?$Total?"
Not the direct quertion but replace
Plc_hldr1=`head -$i place_holder.txt | tail -1 | awk -F ' ' '{ print $1 }'`
Plc_hldr2=`head -$i place_holder.txt | tail -1 | awk -F ' ' '{ print $2 }'`
by
Plc_hldr1=`sed -n "$i {s/ .*//p;q}"`
Plc_hldr2=`sed -n "$i {s/[^ ]\{1,\} \{1,\}\([^ ]\{1,\}\) .*/\1/p;q}"`
and with aix/ksh
sed -n "$i {s/\([^ ]\{1,\} \{1,\}[^ ]\{1,\}\) .*/\1/p;q}" | read Plc_hldr1 Plc_hldr2

Print out onto same line with ":" separating variables

I have the following piece of code and would like to display HOST and RESULT side by side with a : separating them.
HOST=`grep pers results.txt | cut -d':' -f2 | awk '{print $1}'`
RESULT=`grep cleanup results.txt | cut -d':' -f2 | awk '{print $1}' | sed -e 's/K/000/' -'s/M/000000/'`
echo ${HOST}${RESULT}
Please can anyone assist with the final command to display these, I am just getting all of hosts and then all of results.
You probably want this:
HOST=( `grep pers results.txt | cut -d':' -f2 | awk '{ print $1 }'` ) #keep the output of the command in an array
RESULT=( `grep cleanup results.txt | cut -d':' -f2 | awk '{ print $1 }' | sed -e 's/K/000/' -'s/M/000000/'` )
for i in "${!HOST[#]}"; do
echo "${HOST[$i]}:${RESULT[$i]}"
done
A version that works without arrays, using an extra file handle to read from 2 sources at at time.
while read host; read result <&3; do
echo "$host:$result"
done < <( grep peers results.txt | cut -d: -f2 | awk '{print $1}' ) \
3< <( grep cleanup results.txt | cut -d':' -f2 | awk '{print $1}' | sed -e 's/K/000/' -'s/M/000000/')
It's still not quite POSIX, as it requires process substitution. You could instead use explicit fifes. (Also, an attempt to shorten the pipelines that produce the hosts and results. It's probably possible to combine this into a single awk command, since you can either do the substitution in awk, or pipe to sed from within awk. But this is all off-topic, so I leave it as an exercise to the reader.)
mkfifo hostsrc
mkfifo resultsrc
awk -F: '/peers/ {split($2, a, ' '); print a[1]}' results.txt > hostsrc &
awk -F: '/cleanup/ {split($2, a, ' '); print a[1]}' results.txt | sed -e 's/K/000' -e 's/M/000000/' > resultsrc &
while read host; read result <&3; do
echo "$host:$result"
done < hostsrc 3< resultsrc

Bash: "xargs cat", adding newlines after each file

I'm using a few commands to cat a few files, like this:
cat somefile | grep example | awk -F '"' '{ print $2 }' | xargs cat
It nearly works, but my issue is that I'd like to add a newline after each file.
Can this be done in a one liner?
(surely I can create a new script or a function that does cat and then echo -n but I was wondering if this could be solved in another way)
cat somefile | grep example | awk -F '"' '{ print $2 }' | while read file; do cat $file; echo ""; done
Using GNU Parallel http://www.gnu.org/software/parallel/ it may be even faster (depending on your system):
cat somefile | grep example | awk -F '"' '{ print $2 }' | parallel "cat {}; echo"
awk -F '"' '/example/{ system("cat " $2 };printf "\n"}' somefile

Resources