[SOLVED] Bash

deepakdeore2004 · 07-26-2013, 05:56 AM

How do I sort the data which is on line line.

eg. one line contains below data

web2 web5 web3 web1 web6

I want it to be sorted like

web1 web2 web3 web5 web6

sort command doesnt have the option for this requirement.

H_TeXMeX_H · 07-26-2013, 05:59 AM

Indeed, sort can only sort lines, so you have to arrange them as lines:

Code:

bash-4.2$ echo 'web2 web5 web3 web1 web6' | tr ' ' '\n'
web2
web5
web3
web1
web6
bash-4.2$ echo 'web2 web5 web3 web1 web6' | tr ' ' '\n' | sort
web1
web2
web3
web5
web6
bash-4.2$ echo 'web2 web5 web3 web1 web6' | tr ' ' '\n' | sort | tr '\n' ' '
web1 web2 web3 web5 web6 bash-4.2$

deepakdeore2004 · 07-26-2013, 06:20 AM

Great, I didn't think about this way.
Thx.

danielbmartin · 07-26-2013, 07:28 AM

With this InFile ...

Code:

now is the time
for all good men
to come to the aid
of their party

... this code ...

Code:

awk '{split($0,w); s="";
  for (i=1;i<=asort(w);i++) s=s w[i] " "; print s}'  \
  $InFile >$OutFile

...produced this OutFile ...

Code:

is now the time 
all for good men 
aid come the to to 
of party their

Note this minor glitch: each line has a trailing blank.

Daniel B. Martin

reza_zah1991 · 07-26-2013, 09:19 AM

awk is huge , heavy

it seems to ride to home with tractor,,

tractor & bicycle do that work for you but tractor have more price

you can do everything awk can with ' cut+sed '

danielbmartin · 07-26-2013, 09:46 AM

Quote:

Originally Posted by reza_zah1991

you can do everything awk can with 'cut+sed '

The task presented by OP is to sort words within each line. If you can do this with cut and sed please post the code.

Daniel B. Martin

H_TeXMeX_H · 07-26-2013, 09:56 AM

The awk solution is a good alternative. There are many ways to do it.

ntubski · 07-26-2013, 02:32 PM

Quote:

Originally Posted by reza_zah1991

you can do everything awk can with ' cut+sed '

Quote:

Originally Posted by Alan Perlis

Beware of the Turing tar-pit in which everything is possible but nothing of interest is easy.

I think sorting data in sed would be ridiculously difficult. Sorting is not part of standard awk (the given awk solution requires GNU awk), but if you felt like restricting yourself to basic tools, it would be not be nearly as hard to implement in awk as compared to sed.

@danielbmartin: I think your solution calls asort() once for every word in the line, for efficiency (only relevant if you have really long lines) one might want to move that call before the loop.

danielbmartin · 07-26-2013, 05:18 PM

Quote:

Originally Posted by ntubski

@danielbmartin: I think your solution calls asort() once for every word in the line, for efficiency (only relevant if you have really long lines) one might want to move that call before the loop.

My earlier post failed to give a deserved attribution. The awk solution was originally posted by firstfire on 02-02-12. I simplified it, only a bit.

http://www.linuxquestions.org/questi...h-line-927306/

If you can improve the efficiency please post your version of the code.

Daniel B. Martin

ntubski · 07-26-2013, 06:19 PM

To demonstrate the efficiency difference, I will pass input with a pathologically long line:

Code:

~/tmp$ time yes | head -1000 | tr '\n' ' ' | gawk '{ split($0, w); s = "";
  for (i=1; i <= asort(w); i++) 
     s = (s w[i] " ");
  print s}' >/dev/null
> > > 
real	0m2.593s
user	0m2.568s
sys	0m0.008s
~/tmp$ time yes | head -1000 | tr '\n' ' ' | gawk '{ split($0, w); s = ""; wc = asort(w);
  for (i=1; i <= wc; i++) 
     s = (s w[i] " ");
  print s}' >/dev/null
> > > 
real	0m0.037s
user	0m0.028s
sys	0m0.004s

konsolebox · 07-26-2013, 09:52 PM

First time I actually tried to read the manual of gawk, one of the things I always remember is how asort is not implemented in ordinary awk. The new variation of it asorti is even also a newcomer. Still sorting in awk is easy to implement like this:
http://code.google.com/p/lawker/sour.../quicksort.awk
And mergesort fans could have a little more complex solutions: http://code.google.com/p/andorian-bl...merge_sort.awk

danielbmartin · 07-27-2013, 10:10 AM

Quote:

Originally Posted by ntubski

Code:

gawk '{ split($0, w); s = ""; wc = asort(w);
  for (i=1; i <= wc; i++) 
     s = (s w[i] " ");
  print s}' >/dev/null

Excellent! Thank you!

Daniel B. Martin