UNIX for Dummies Questions & Answers

hello!
I am trying to use sed to copy specific set of lines from a file for which the starting and ending line numbers of the lines to be copied are stored in shell variables. How can i copy those lines?

if the input_file is something like this
and if the following is the script

Code:

a=2
b=4
sed '/$a/,/$b/p' input_file > new_file

the output required is
the shell variable a and b can presume different values accordingly. how can these shell variables be used in sed ? Can awk be used for this?

use double quote instead of simple one
and remove the slash by the way :

the more simple:

Code:

sed $a,$b'!d' infile

but those may also work:

Code:

sed ''$a,$b'!d' infile

(4 simple quotes) or

Code:

sed "$a,$b"'!d' infile

(2 double quotes 2 simple quotes)

or simply

Code:

sed -n $a,$b'p' infile

hey! ctsgnb
None of the modified above suggestions worked for me..
however, the earlier solution suggested by you.. i.e

Code:

sed "/$a/,/$b/!d" input_file > new_file

worked perfectly for me. Thank you, very much.

All depends what you put in $a and $b :

if you use line number, you should not use the slash

Code:

a=2
b=4
sed $a,$b'!d' infile

but if you want to do your selection based on the pattern matching, then you need the slash

Code:

a="def"
b="jkl"
sed "/$a/,/$b/!d" infile

so your example was confusing because you stated :

Code:

a=2
b=4

so i assumed you were working with line number instead of pattern.

ok! i now get it! what you are saying is exactly correct!

For the line numbers as in the example given by me.. the slash is not required and for matching the patterns it is required (helped me in the other way).

Thank you so much for the clarification.