I am trying to look for a way to delete every 3rd and 5th line but not the 15th using sed, but this is the thing: you can't make use of the ~ way (GNU). It has to be something like
sed 'n;n;d' test
but I can't figure out how to combine the 3...
Example input
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Example output:
1
2
4
7
8
11
13
14
15
It'll need to be in sed, no awk or perl
awk command is easier to understand for this requirement:
awk 'NR==15 || (NR%3 && NR%5)' file
1
2
4
7
8
11
13
14
15
ugh:
$ seq 15 | sed -n 'p;n;p;n;n;p;n;n;n;p;n;p;n;n;n;p;n;n;p;n;p;n;p'
1
2
4
7
8
11
13
14
15
This might work for you (GNU sed):
sed '0~15b;0~3d;0~5d' file
Using gnu sed:
sed '15p;0~3d;0~5d' test
here is the test result from above awk/sed commands:
seq 99 |awk 'NR==15 || (NR%3 && NR%5)' > anubhava.txt
seq 99 |sed -n 'p;n;p;n;n;p;n;n;n;p;n;p;n;n;n;p;n;n;p;n;p;n;p' > glenn.jackman.txt
seq 99 |sed '0~15b;0~3d;0~5d' > potong.txt
seq 99 |sed '15p;0~3d;0~5d' > bmw.txt
$diff anubhava.txt glenn.jackman.txt
17a18
> 30
25a27
> 45
33a36
> 60
41a45
> 75
49a54
> 90
$ diff -q anubhava.txt potong.txt
Files anubhava.txt and potong.txt differ # same problem that can't delete line 30, 45, 60, etc.
$ diff -q anubhava.txt bmw.txt
$
Related
I have a file which has n number of rows, i want it's data to be distributed in 7 files as per below order
** my input file has n number of rows, this is just an example.
Input file
1
2
3
4
5
6
7
8
9
10
11
12
13
14
1
5
16
17
.
.
28
Output file
1 2 3 4 5 6 7
14 13 12 11 10 9 8
15 16 17 18 19 20 21
28 27 26 25 24 23 22
so if i open the first file it should have rows
1
14
15
28
similarly if i open the second file it should have rows
2
13
16
27
similarly output for the other files as well.
Can anybody please help, with below code it is doing what is required but not in required order.
awk '{print > ("te1234"++c".txt");c=(NR%n)?c:0}' n=7 test6.txt
1 2 3 4 5 6 7
8 9 10 11 12 13 14
15 16 17 18 19 20 21
22 23 24 25 26 27 28
EDIT: Since OP has changed sample of Input_file totally different so adding this solution now, again this is written and tested with shown samples only.
With xargs + single awk: (recommended one)
xargs -n7 < Input_file |
awk '
FNR%2!=0{
for(i=1;i<=NF;i++){
print $i >> (i".txt")
close(i".txt")
}
next
}
FNR%2==0{
for(i=NF;i>0;i--){
count++
print $i >> (count".txt")
close(i".txt")
}
count=""
}'
Initial solution:
xargs -n7 < Input_file |
awk '
FNR%2==0{
for(i=NF;i>0;i--){
val=(val?val OFS:"")$i
}
$i=val
val=""
}
1' |
awk '
{
for(i=1;i<=NF;i++){
print $i >> (i".txt")
close(i".txt")
}
}'
Above could be done with single awk too will add xargs + awk(single) solution in few mins too.
Could you please try following, written and tested with shown samples in GNU awk.
awk '{for(i=1;i<=NF;i++){print $i >> (i".txt");close(i".txt")}}' Input_file
The output file counter could descend for each second group of seven:
awk 'FNR%n==1 {asc=!asc}
{
out="te1234" (asc ? ++c : c--) ".txt";
print >> out;
close(out)
}' n=7 test6.txt
$ ls
file tst.awk
$ cat tst.awk
{ rec = (cnt % 2 ? $1 sep rec : rec sep $1); sep=FS }
!(NR%n) {
++cnt
nf = split(rec,flds)
for (i=1; i<=nf; i++) {
out = "te1234" i ".txt"
print flds[i] >> out
close(out)
}
rec=sep=""
}
.
$ awk -v n=7 -f tst.awk file
.
$ ls
file te12342.txt te12344.txt te12346.txt tst.awk
te12341.txt te12343.txt te12345.txt te12347.txt
$ cat te12341.txt
1
14
15
28
$ cat te12342.txt
2
13
16
27
If you can have input that's not an exact multiple of n then move the code that's currently in the !(NR%n) block into a function and call that function there and in an END section.
This might work for you (GNU sed & parallel):
parallel 'echo {1}~14w file{1}; echo {2}~14w file{1}' ::: {1..7} :::+ {14..8} |
sed -n -f - file &&
paste file{1..7}
Create a sed script to write files named filen where n is 1 thru 7 (see above first set of parameters in the parallel command and also in the paste command).
The sed script uses the n~m address where n is the starting address and m is the modulo thereafter.
The distributed files are created first and the paste command then joins them all together to produce a single output file (tab separated by default, use paste -d option to get desired delimiter).
Alternative using Bash & sed:
for ((n=1,m=14;n<=7;n++,m--));do echo "$n~14w file$n";echo "$m~14w file$n";done |
sed -nf - file &&
paste file{1..7}
I'm trying to print text from line 10 to 20 and then reverse their positions.
I've tried this:
sed '10!G;h;$!d' file.txt
But it only prints from 10 to end of the file. Is there any way to stop it at line 20 by using only one sed command?
Almost there, you just need to replace $!d with the 'until' line-number
sed -n '10,20p' tst.txt
// Prints line 10 <--> 20
sed -n '10!G;h;20p' tst.txt
// Prints REVERSE line 10 <--> 20
output:
20
19
18
17
16
15
14
13
12
11
10
tst.txt:
1
2
3
4
...
19
20
Info
You can use this to print a range of lines:
sed -n -e 10,20p file.txt | tac
tac will reverse the order of the lines
And for those of you without tac (like those mac users out there):
sed -n -e 10,20p file.txt | tail -r
I have the following test file:
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
I want to separate it in a way that each file contains the last line of the previous file as the first line. The example would be:
file 1:
1
2
3
4
5
file2:
5
6
7
8
9
file3:
9
10
11
12
13
file4:
13
14
15
16
17
file5:
17
18
19
20
That would make 4 files with 5 lines and 1 file with 4 lines.
As a first step, I tried to test the following commands I wrote to get only the first file which contains the first 5 lines. I can't figure out why the awk command in the if statement, instead of printing the first 5 lines, it prints the whole 20?
d=$(wc test)
a=$(echo $d | cut -f1 -d " ")
lines=$(echo $a/5 | bc -l)
integer=$(echo $lines | cut -f1 -d ".")
for i in $(seq 1 $integer); do
start=$(echo $i*5 | bc -l)
var=$((var+=1))
echo start $start
echo $var
if [[ $var = 1 ]]; then
awk 'NR<=$start' test
fi
done
Thanks!
Why not just use the split util available from your POSIX toolkit. It has an option to split on number of lines which you can give it as 5
split -l 5 input-file
From the man split page,
-l, --lines=NUMBER
put NUMBER lines/records per output file
Note that, -l is POSIX compliant also.
$ ls
$
$ seq 20 | awk 'NR%4==1{ if (out) { print > out; close(out) } out="file"++c } {print > out}'
$
$ ls
file1 file2 file3 file4 file5
.
$ cat file1
1
2
3
4
5
$ cat file2
5
6
7
8
9
$ cat file3
9
10
11
12
13
$ cat file4
13
14
15
16
17
$ cat file5
17
18
19
20
If you're ever tempted to use a shell loop to manipulate text again, make sure to read https://unix.stackexchange.com/questions/169716/why-is-using-a-shell-loop-to-process-text-considered-bad-practice first to understand at least some of the reasons to use awk instead. To learn awk, get the book Effective Awk Programming, 4th Edition, by Arnold Robbins.
oh. and wrt why your awk command awk 'NR<=$start' test didn't work - awk is not shell, it has no more access to shell variables (or vice-versa) than a C program does. To init an awk variable named awkstart with the value of a shell variable named start and then use that awk variable in your script you'd do awk -v awkstart="$start" 'NR<=awkstart' test. The awk variable can also be named start or anything else sensible - it is completely unrelated to the name of the shell variable.
You could improve your code by removing the unneccesary echo cut and bc and do it like this
#!/bin/bash
for i in $(seq $(wc -l < test) ); do
(( i % 4 != 1 )) && continue
tail +$i test | head -5 > "file$(( 1+i/4 ))"
done
But still the awk solution is much better. Reading the file only once and taking actions based on readily available information (like the linenumber) is the way to go. In shell you have to count the lines, there is no way around it. awk will give you that (and a lot of other things) for free.
Use split:
$ seq 20 | split -l 5
$ for fn in x*; do echo "$fn"; cat "$fn"; done
xaa
1
2
3
4
5
xab
6
7
8
9
10
xac
11
12
13
14
15
xad
16
17
18
19
20
Or, if you have a file:
$ split -l test_file
I have a file/string containing the following:
[1-9]
[11-12]
[10-15]
I then want to expand that to become this:
1 2 3 4 5 6 7 8 9
11 12
10 11 12 13 14 15
I know how to do it in a very long way (first capture the two numbers and then expand them using a for loop).
I would like to know if there is a faster/smarter way of achieve the same.
One way:(Pure bash solution)
while IFS=- read l1 l2
do
eval echo ${l1/[/{}".."${l2/]/}}
done < file
There are several solutions.
Solution 1:
sed 's/^/echo /; s/[[]/{/; s/]/}/; s/-/../' | bash
Example:
$ cat 1.txt | sed 's/^/echo /; s/[[]/{/; s/]/}/; s/-/../' | bash
1 2 3 4 5 6 7 8 9
11 12
10 11 12 13 14 15
Solution 2:
tr '[]-' ' ' | sed "s/^/seq -s' '/" | bash
Example:
$ cat 1.txt | tr '[]-' ' ' | sed "s/^/seq -s' '/" | bash
1 2 3 4 5 6 7 8 9
11 12
10 11 12 13 14 15
If you're confident that your input all matches that pattern:
while read a; do
seq -s' ' $(echo "$a" | tr '[]-' ' ')
done
Add error checking as appropriate.
Here's a one-liner:
cat lines | sed -E -e 's/\[|]//g' -e 's/-/ /g' | xargs -n 2 seq -s ' ' -t '\n'
As in:
$ cat <<EOF | sed -E -e 's/\[|]//g' -e 's/-/ /g' | xargs -n 2 seq -s ' ' -t '\n'
> [1-9]
> [11-12]
> [10-15]
> EOF
1 2 3 4 5 6 7 8 9
11 12
10 11 12 13 14 15
I want to use a command to make a crontab that plays an alarm (for my wife). The program is called ipraytime and it gives an output like this.
$ ipraytime -u +2
Prayer schedule for,
City : Custom
Latitude : 021�� 25' 12" N
Longitude : 039�� 49' 47" E
Angle Method : Umm Al-Qurra University
TimeZone : UTC+2.0
Qibla : 061�� 45' 42" W of true North
Date Fajr Shorooq Zuhr Asr Maghrib Isha
--------------------------------------------------------------------
[09-05-2012] 4:19 5:43 12:16 15:35 18:48 20:18
Today's Imsaak : 4:11
Tomorrow's Imsaak : 4:10
Tomorrow's Fajr : 4:18
What i want is that the times format good for a crontab which means i need to switch places of the minute and hour. To be 19 4 instead.
I have made this command but don't know how to make that switch.
ipraytime -u +2| awk 'NR==12 {print $2"\n"$3"\n"$4"\n"$5"\n"$6"\n"$7}' | sed 's/:/ /g'
This gives me an output like this
4 19
5 43
12 16
15 35
18 48
20 18
But i want it to be like this
19 4
43 5
16 12
35 15
48 18
18 20
As that is what a crontab is using. I have played with sort a bit but couldn't find a solution there either.
(Sorry for the bad topic.. didn't know how to write a good one for this)
It's not necessary to use sed at all.
$ ipraytime -u +2 | awk -F ' +|:' 'NR == 12 {for (i = 2; i <= 12; i += 2) print $(i+1), $i}'
19 4
43 5
16 12
35 15
48 18
18 20
Use sed 's/\(.*\):\(.*\)/\2 \1/'
Command:
ipraytime -u +2 | awk 'NR==12 {print $2"\n"$3"\n"$4"\n"$5"\n"$6"\n"$7}'
| sed 's/\(.*\):\(.*\)/\2 \1/'