shell scripe replace lines in file.

nenad_m · 11-27-2008, 04:55 AM

Hi,

I have a file with following content:

b = 5
b = 1
b = 0
b = 13
b = 245
b = 242
stat = invalid
b = 4
b = 8
b = 0
b = 1
b = 53
stat = valid
...
b = 6
b = 8
b = 0
b = 17
b = 70
b = 113
b = 200
stat = invalid

I need file that looks like:
5, 1, 0, 13, 245, 242, invalid
4, 8, 0, 1, 53, valid
...
6, 8, 0, 17, 70, 113, 200, invalid

How to do that?

colucix · 11-27-2008, 05:14 AM

What have you tried so far?

nenad_m · 11-27-2008, 10:18 AM

Well i succeed to remove characters before numbers and word 'stat', but i don't know how to group multiple rows in one row.

colucix · 11-27-2008, 10:44 AM

Well, using shell scripting you can simply concatenate values by putting them one after another, followed by a comma. For example, every time you read a value in a row, you can append this value to a string, then put a comma, then add the new value and so on. Here is an example of what I'm talking about:

Code:

#!/bin/bash
while read line
do
  #if echo $line | grep -q stat   # for bash version < 3
  if [[ "$line" =~ "stat" ]]      # for bash version >= 3
  then
    echo -n $number
    echo $(echo $line | cut -d ' ' -f3)
    unset number
  else
    if [ -z $number ]
    then
      number=$(echo $line | cut -d ' ' -f3),
    else
      number=$number$(echo $line | cut -d ' ' -f3),
    fi
  fi
done < testfile   # the name of your input file here

Here, you read a "line", extract the "number" and append it to the previous number (if any) followed by a comma. When you encounter a line containing the string "stat", you echo the "valid/invalid" part, unset the previous "number" and start a new line of output. Hope this helps.

colucix · 11-27-2008, 10:46 AM

And here is an awk solution very similar to the previous one for the logic but using arrays:

Code:

awk '
! /stat/ {
  array[count++] = $3
}
/stat/ {
  for ( i=0; i<=count-1; i++ ){
    printf "%s,",array[i]
    delete array[i]
  }
  print $3
  count = 0
} ' testfile

radoulov · 11-27-2008, 10:59 AM

And another one:

Code:

awk -F' = ' 'ORS=/stat/?RS:OFS{$0=$NF}1' OFS=, infile

Or this (if you want the space after the comma):

Code:

awk -F' = ' 'ORS=/stat/?RS:OFS{$0=$NF}1' OFS=', ' infile

With Perl:

Code:

perl -ne'print/= (.*)/,/stat/?$/:", "' infile