Shell Programming and Scripting

I am trying to understand what the grep command in ubuntu is trying to do here.
The contents of my test file is given below

Code:

harsha@harsha-H67MA-USB3-B3:~/Documents$ cat data
abcd
efghi
jklmno
pqr
stuv
wxyz

When I grep for 3 dots (...) without the parenthesis as follows I would expect the output to match the first 3 characters in the above file. But the output I am seeing shows that it matches more than 3 characters in some lines. Can anyone help me here understand why the 3rd line is being shown to match more than 3 characters.

Code:

harsha@harsha-H67MA-USB3-B3:~/Documents$ grep ... data
abcd
efghi
jklmno
pqr
stuv
wxyz

Grep 3 char from beginning... would this solve your problem ?

Code:

$ grep '^...' data

OR

$ grep -E '^.{3}' data

abcd
efghi
jklmno
pqr
stuv
wxyz

Yes it solves.. but why does'nt

Code:

...

work only for the 3rd line

just ... match every 3 suppose if your line contains 6 or 9 char string then also it matches, 3 char as one set.

Execute this on your terminal, you will come to know

Code:

# considers 5 digits pairs  
$ echo '1234567.1234567' | grep -E '[[:digit:]]{5}'
1234567.1234567

# considers 5 digits from beginning 
$ echo '1234567.1234567' | grep -E '^[[:digit:]]{5}'
1234567.1234567

If you are looking for just three character on a line (not zero, one or two & not four or more) then you need to mark the beginning and end of the line in your expression. The carat ^ marks the beggining and the dollar $ is the end of a record. With these you can set the expression and search exactly:-

Code:

grep "^...$" data

Using grep is great, but you have to be strict with your expression if that's what you need. You have to refine your expression(s) so that:-
There is a good book by O'Reilly for sed & awk which has good explanations for expressions. It is worth a read. You don't have to get full into sed & awk but it's great to know as it will help in so many ways.




Robin