How to extract text blocks seperated by blank line

art84_LV · 11-12-2009, 06:38 AM

Hi,

I have file which contains information written in blocks (every block is different). Is it possible to read every block one by one to another file (one block per file).

The input is something like this

<block1>
<empty line>
<block2>
<empty line>
...
...
...
<block25>
<empty line>
<block26>

I need to read <block1> to file.tmp. Make some manipulations with it and then rewrite the contents of file.tmp with <block2> make some manipulations with content of file.tmp and rewrite file.tmp with <block3> and initiate some manipulations with it. And so on....

Any ideas?

pixellany · 11-12-2009, 06:53 AM

Welcome to LQ!!

Please give a bit more detail on what you are trying to do. e.g. why move blocks to a temp file---why not do the manipulations in place?

art84_LV · 11-12-2009, 07:15 AM

Quote:

Originally Posted by pixellany

Welcome to LQ!!

Please give a bit more detail on what you are trying to do. e.g. why move blocks to a temp file---why not do the manipulations in place?

Every block can contain different information (block formatting and line count is different too). The reason why I want to operate with each block is to run some if statements for each block to check the format and according on that grep needed info.

An example

BLOCK_1
Item: $item1
unnecessary info
Time: $time1
unnecessary info
unnecessary info
Type: $type1
unnecessary info
$name1
unnecessary info
unnecessary info
unnecessary info
<blank line>
BLOCK_2
Item: $item2
Time: $time2
unnecessary info
Type: $type2
unnecessary info
unnecessary info
unnecessary info
{$destination} # in case of $destination {} are used
unnecessary info
<blank line>

OUTPUT FILE:
$time1 $item1 $type1 $name
$time2 $item2 $type2 $destination

As you see $time, $item and $name can be extracted using simple grep (grep Time, grep Item, grep Type) the problem is with extracting other variables.

ghostdog74 · 11-12-2009, 07:40 AM

use gawk

Code:

awk -vRS= -vFS="\n" '{
  for(i=1;i<=NF;i++){
    gsub(/\#.*/,"",$i)
    if ( $i~/item|time|type|name|destination/){
        printf $i" "
    }
  }
  print ""  
}' file

output

Code:

$ ./shell.sh
$item1 $time1 $type1 $name1
$item2 $time2 $type2 {$destination}

art84_LV · 11-12-2009, 08:03 AM

Quote:

Originally Posted by ghostdog74

use gawk

Code:

awk -vRS= -vFS="\n" '{
  for(i=1;i<=NF;i++){
    gsub(/\#.*/,"",$i)
    if ( $i~/item|time|type|name|destination/){
        printf $i" "
    }
  }
  print ""  
}' file

output

Code:

$ ./shell.sh
$item1 $time1 $type1 $name1
$item2 $time2 $type2 {$destination}

It works, GREAT, thanks a lot.

Could you please explain what you have written

I am not familiar with gawk, but I really would like to learn it. Maybe you could provide some useful tutorials?

ghostdog74 · 11-12-2009, 08:15 AM

Quote:

Originally Posted by art84_LV

but I really would like to learn it. Maybe you could provide some useful tutorials?

see my sig for gawk link.

pixellany · 11-12-2009, 11:22 AM

and of course one of the best tutorial sites around:
http://www.grymoire.com/Unix/

art84_LV · 11-16-2009, 08:17 AM

Guys, Thanks a lot for help. But I have another question. Tried to resolve it by my own, but unfortunately without success. With your script the output is the following ( _ means space ) :

Item___: $item1 $time = 2009.15.11 9:42:15.397927000 Type___: $type1 $name

And I want to make it as following:

$time = 9:42:15 Item: $item1 Type: $type1 $name

Is it possible to make such kind of formatting in this script? If it is not possible inside the script, what workaround should I use to make this thing work?

ghostdog74 · 11-16-2009, 05:36 PM

show what you have tried.

art84_LV · 11-17-2009, 01:02 AM

Quote:

Originally Posted by ghostdog74

show what you have tried.

I don't store unsuccessful jobs. I have tried to manipulate output format in printf line

))