Shell Programming and Scripting

Hi All,

I am new to shell script. And right now learning regular expression. I am using BASH as a shell and RedHat Linux. I have couple of doubts regarding the regular expression

1) +
This is matches one or more occurrence of the preceding pattern.
So if the file contains following data
file.txt
a
aa
aba
bbb
ccc

Code:

grep -i 'a+' file.txt

should print following
a
aa
aba

But to print above I need to have

Code:

 grep -i a\+' file.txt

2) ?
should match 0 or 1 occurrence of the preceding pattern

But when I use

Code:

 grep -i 'a?' file.txt

it pritnts nothing.. And when I use

Code:

grep -i 'a\?' file.txt

it pritints all the data in the file. i.e output was as below
a
aa
aba
bbb
ccc

I was expecting the output as below
a
bbb
ccc

Now I have a doubt 0 or 1 means exactly 0 or 1 ?? or is it mean 0 or more then 0 ??? if it means 0 or more then 0 then whats the differnece between * (asterisk ) and ? ( question mark )??


3)

Code:

grep -i 'a\{1,\}' file.txt

and

Code:

grep -i 'a\{1\}' file.txt

prints the same data as below

a
aa
aba

I was expecting this output only when I use

Code:

grep -i 'a\{1,\}' file.txt

and should print below line when I use

Code:

grep -i 'a\{1\}' file.txt

a

Please Please help me.........
Thnx in advance

Problem is, that grep just understands very limited regular expressions until you add in the switch "-E" or use "egrep" and you have to escape a + of course so it isn't interpreted as "just another character":

Code:

root@isau02:/data/tmp/testfeld> cat infile
a
aa
aba
bbb
ccc
root@isau02:/data/tmp/testfeld> grep -E a\+ infile
a
aa
aba

For 2):
A single character is a dot, not a question mark.

For 3):
{1} also stands for minimum. It's the same as {1,}. And even {1,4} will print out every line where at least one "a" is in. There is a very good pocke guide on awk and sed from O'Reilly. They have a good overview on regular expressions in the first few pages, very helpful.

Hi
Thnx for your reply
.
1) But when I refer the book and some other on line articles they says that "?" will print 0 or 1 occurrence of the preceding character ?? ( I was understanding it as exact 0 or exact 1, my mistake )

But now it seems that "*" and "?" are same. or is there any difference ??

2) And what should I do if I need to match the exact number of occurrence in the string.

i.e. I need to find the string that contains only 2 "a" ( not less then 1 and not more then 2 ).

Try:

Code:

grep -i "a.*a[^a]*" file.txt

hey thanx,

But above code prints all the strings that contains 2 or more then 2 "a" characters.

I want strings that contain exactly 2 "a" charcters.

for example if file contains
a
aa
aaa
abab
ababab

I want only following string.
aa ( only 2 "a" )
abab ( only 2 "a" )

But you code prints the following

aa
aaa ( I don't want it as it contains the 3 "a" )
abab
ababab ( I don't want it as it contains the 3 "a" )

I understood what you want, but to explain the ?, I read about it and fooled around, which might explain it a bit:

Code:

echo -e "November" | egrep Nov\(ember\)\?
November
echo -e "Nov" | egrep Nov\(ember\)\?
Nov
echo -e "Noveember" | egrep Nov\(ember\)\?
Noveember
echo -e "No" | egrep Nov\(ember\)\?
echo -e "Novrebm" | egrep Nov\(ember\)\?
Novrebm

So it is as the description says, none or one occurence. Here it is a group of characters, but they interpret as single characters, not as a complete string.

So much for that.

A quick workaround:

Code:

> cat text
a
aa
aaa
abab
ababab
jjhk
jjsasa
jjsasasda
> awk 'NF==3' FS='a' text
aa
abab
jjsasa