Bdtyktl

I have a FILE1

NAMES START



Hi, How are you?

Good to see you



start

Aaron

Kyle

Robert

stop



Official use only

Stamps here



start

Riley

Gayle

stop



Bicycles here

United Pawns



start

Alex

Ford

Sergio

stop



NAMES STOP

Here's what I wanna do:

If "NAMES START" string is present, transfer the contents each "start" and "stop", to a new FILE2, leaving out the start and stop themselves in the new FILE2.

So FILE2 should look like this:

Aaron

Kyle

Robert

Riley

Gayle

Alex

Ford

Sergio

Help please!

Getting content between named blocks of text

Lets address getting the contents inside start..stop block because that's the major task.

grep is line-matching tool, hence matching across multiple lines is difficult and grep is usually not used for such task, though it's not impossible. However, best approach in such case is to use awk to get patterns between specific lines, and then filter out the start and stop flags:

$ awk '/start/,/stop/' input.txt | grep -v 'start|stop'

Aaron

Kyle

Robert

Riley

Gayle

Alex

Ford

Sergio

If we want to get rid of pipeline, we can do:

# {print} is actually unnecessary, as default for awk is to print if

# expression evaluates to true, so it's enough to have

# awk '/start/{flag=1;next};/stop/{flag=0};flag' input.txt

$ awk '/start/{flag=1;next};/stop/{flag=0};flag == 1 {print}' input.txt

Aaron

Kyle

Robert

Riley

Gayle

Alex

Ford

Sergio

There's of course other methods, such as sed or perl. As for grep, it's often suggested to use grep -Pzo flags, however probably due to multiple occurrences of start..block it doesn't work properly ( only one match returned):

$ grep -zPo --color 'startn(.*n.*)nstop' input.txt

start

Riley

Gayle

stop

For the sake of simplicity (though arguably awk example is the simplest) and to avoid dealing with regex patterns, we can also resort to basic scripting:

#!/bin/bash



printline=0

while IFS= read -r line; do



    # set flag for printing or not printing

    case $line in 

        "start") printline=1; continue;;

        "stop")  printline=0; continue;;

    esac



    # check the flag

    if [ "$printline" -eq 1  ]; then

        printf "%sn" "$line"

    fi



# first positional parameter is the file we read

done < "$1"

And here's how the script works:

$ chmod +x readblocks.sh

$ ./readblocks.sh input.txt

Aaron

Kyle

Robert

Riley

Gayle

Alex

Ford

Sergio

Checking for NAMES START and transfering

If "NAMES START" string is present, transfer the contents each "start" and "stop", to a new FILE2, leaving out the start and stop themselves in the new FILE2.

Well, that's just grep 'NAMES START' input.txt. So we can check for that via

if grep -q 'NAMES START' input.txt; then

    # whatever method you like goes here

fi

Looking at your example, NAMES START is the first line of the file. So we can also check for that - check first line as we read the file, instead of opening the file inside if statement suggested above.

Transferring the contents to FILE2 - that's just adding > FILE2.txt to the original command or script you use.

With these suggestions, awk command becomes:

$ awk 'NR==1 && $0 != "NAMES START"{exit};/start/{flag=1;next};/stop/{flag=0};flag' input.txt > FILE2.txt

And the script:

#!/bin/bash



printline=0

linecounter=0

while IFS= read -r line; do

    linecounter=$((linecounter+1))



    case "$line" in

        "NAMES START") continue;;

         *) exit 1;

    esac





    # set flag for printing or not printing

    case $line in 

        "start") printline=1; continue;;

        "stop")  printline=0; continue;;

    esac



    # check the flag

    if [ "$printline" -eq 1  ]; then

        printf "%sn" "$line"

    fi



# first positional parameter is the file we read

done < "$1" > FILE2.txt