Shell Programming and Scripting

Hello,

I have a file temp.txt:
-------------------------
HELLO WORLD
This is a temp file.
TENCHARSHEre
no beginning UPPERCHARS
HI There
-------------------------

What is a sed egrep command that will target lines that begin with 3-10 uppercase chars, and output the first 2 chars?
desired result:
---
HE
TE
---

So far I have been able to extract the targeted lines in egrep using the expression: egrep ^[A-Z]\{3,10\} < temp.txt . . but no luck in sed.

Thank you

Code:

sed -n 's/^\([A-Z][A-Z]\)[A-Z].*/\1/p' temp.txt

Thank you for the reply..
However I have to do the following using egrep.

I tried translating what you gave using -r:
sed -r 's/^([A-Z][A-Z])[A-Z].*/\1/g' temp.txt

Updated temp.txt:
-------------------------
HELLO WORLD
This is a temp file.
TENCHARSHEre
no beginning UPPERCHARS
HI There
ELEVENCHARActers
-------------------------
desired output:
HE
TE

actual output:
HE
This is a temp file.
TE
no beginning UPPERCHARS
HI There
EL

1) How do I ignore irrelevant targets? (i.e. "This is a temp file.", etc..)
2) It also includes 11 upperchars as the target. How do I restrict it to 10 uppercase chars only?

Sed and regex are very difficult concepts for me to grasp, so I appreciate the patience and time. Thank you

Code:

sed -nr '/^[A-Z]{11}/!{s/^([A-Z]{2})[A-Z].*/\1/p;}' temp.txt

Or:

Code:

sed -nr 's/^([A-Z]{2})[A-Z]{1,8}[^A-Z].*/\1/p' temp.txt

Thank you radoulov

The second method is a little more intutitive . . . Is there a way to execute the expression without the -n option?

Thanks again

Why you need to remove the -n option?