Shell Programming and Scripting

How to sort such files which contains records of varying length and varying lines? (With respect to Bash shell)

Eg:
Each record begins with a sting of 1/0(binary) which may or may not be followed by properties like AB,BS etc.
I have to sort such records on the basis of 1/0 string and keep the properties associated with that record intact.

Following lines are contained in the file

10101010101111101
AS sasa
BS kkk
1110000101010110
BS jsa
0110000101010100
1010000101010110
AS asas
BS sasa
CS asa

P.S: 1. A new record starts wherever a new line is starting with 1/0
2. Anyline starting with space is properties associated with the last
1/0 record
3. The number of records depicted in above file is 4




Please help in this regard.

Add a record separator = I used |, then roll the first records and sub-records into one line, sort, then unroll.

Code:

sed 's/^ /|/' filename | \
awk 'BEGIN{ getline; printf "%s", $0}
    { if(substr($0,1,1)=="|"){
         printf "%s", $0
      }
      else {
         printf "\n%s", $0
      }
    }    
     END{printf "\n"} 
     ' > newfile
sort newfile | sed 's/|/\n /g' > sortedfile

input

Code:

10101010101111101
 AS sasa
 BS kkk
1110000101010110
 BS jsa
0110000101010100
1010000101010110
 AS asas
 BS sasa
 CS asa

output

Code:

0110000101010100
1010000101010110
 AS asas
 BS sasa
 CS asa
10101010101111101
 AS sasa
 BS kkk
1110000101010110
 BS jsa

Hi Jim !

Thanks alot...

Please tell me if you have any links for Shell Script beginer's

Regards
Sandeep