Search Results

Now comes the full beauty of the proposal in post#3 into play: if you insert

LINE=${LINE//$'\t\t'/$'\t'x$'\t'}; right after the do, it might fly...

Yes this is a problem for bash. If you have either ksh93 or zsh on your system (if not you can install it) then you can do this:
while IFS=$'\t\t' read -A CU
do
curl -v -b cookie.txt -d...

In bash, if you read straight into an array you do not need to set and unset IFS globally, so RudiC's suggestion modified, becomes:
while IFS=$'\t' read -a CU
do
curl -v -b cookie.txt -d...

With bash, this snippet - although being far from complete and error proof - could point you in the right direction:

OLDIFS="$IFS"
IFS=$'\t'
while read LINE
do CUARR=($LINE)
echo...

Alright, but the input specification keeps changing, are you sure this is that last exception?

Try:

awk -F'[.,:] *' ' # Use dot, command and colon as fields separator
/^[A-E]\./ { ...

The difference is that apart from the capital letters for QUESTION, there is a space after the colon with zero or more spaces behind it...

So, try:
awk -F'[.,:] *' ' # Use dot, command...

Try:
awk -F'[.,:]' ' # Use dot, command and colon as fields separator
/^[A-E]\./ { # Store the choices in array A with index the first field (letter A-E)
i=$1
$1=x...

awk '/begin/{s=""}{s=s ? s ORS $0:$0}/end/{if(s~/pattern1/)print s RS}'


Glad to know that it worked for you..

So here is details

/begin/{s=""} --> Search keyword begin, if keyword found...

sed -n "/begin/,/end/{H;/begin/h;}; /end/{g;/pattern1/p;}" file

Try :

$ awk '/begin/{s=""}{s=s ? s ORS $0:$0}/end/{if(s~/pattern1/)print s RS}' file
begin
def
ghi
pattern1
end

begin
pattern1
jkl
end

As Scrutinizer says, there are some caveats. If you're sure that we are dealing always with paired lines, and /abc/ is always the first in that pair, you might try a getline after an /abc/ match to...

I think you would need to adjust your solution so that x gets erased once a match for a different line gets found and then you would also need to test if x contains something before printing both x...

Actually your first proposal would work once your add/replace some curly brackets:awk '/abc/ {if ($1>10) {print}};/xyz/ {if ($2>5) {print}}' file
12 14 3 20 45 abc
33 19 10 21 49 xyz
55 41 13 62...

That is a little bit more complicated. Try something like:
awk 'NR%2{p=$0; m=$1; next} p~/abc/ && m>10 && /xyz/ && $2>5{print p ORS $0}' file

output:
55 41 13 62 24 abc
12 44 16 54 62 xyz

Hi, are you looking for something like this?

awk '(/abc/ && $1>10) || (/xyz/ && $2>5)' file

Output:
12 14 3 20 45 abc
33 19 10 21 49 xyz
55 41 13 62 24 abc
12 44 16 54 62 xyz

bash-3.00$ ./test.ksh
3.33
5.34
4.17056
bash-3.00$ cat test.ksh
nawk '{print $7}' test.txt | sort -n | head -1 #min
nawk '{print $7}' test.txt | sort -nr | head -1 #max
nawk...

awk 'NR==1{min=$7;max=$7;sum+=$7;next}{if ($7<min) min=$7;if ($7>max) max=$7;sum+=$7}END{print "min: "min", max: "max", avg: "sum/NR}' file

ARGV will have the values *.pdf i.e. a.pdf, b.pdf and c.pdf
so ARGV[1] is a.pdf, ARGV[2] is b.pdf and so on
this is what the code does

1. reads from the file "file1" # while ((getline < file1) >...

The code below will rename filex-1.pdf to filex-1-apple.pdf:


printf '%s\n' *.pdf |
sort -t- -k2n |
awk -F. 'getline f1 < "file1" {
printf "mv -- \"%s\" \"%s-%s\.%s\"\n", \
...

Try this...

#to delete 1 and 3rd <tr>...</tr> pairs
awk '/<tr>/ {i=i+1} {if(i==1 || i==3){next} print}' file
if further if you want to delete say only the second pair, tweak the above code as...

$ lastline=$(( $(wc -l <tst)-20 ))
$ sed '1,'$lastline'!d' tst
1
2
3
4
5
6
7
8
9
10

Use double quotes
sed "$start,$end d" fileOr, if you have metacharacters in there, use single quotes and turn them on and off to get the $start and $end in:

sed ''$start','$end' d' file


The...

awk '{for (i=1;i<=$2;i++) printf "wget http://somewebsite/page=%s&line=%s\n",$1,i}' infile |sh

while read p cnt ; do
for i in `seq $cnt` ; do
echo "wget http://smthn/page=${p}&line=$i"
done
done < input

here you go..
grep ayat >>>>insert file-name here<<<< | sed 's/^.*), //' | sed 's/,.*$//' | nl

cheers