UNIX for Dummies Questions & Answers

Hi, can anyone explain step by step what this line of sed code is trying to do. Thanks.

Sed -e "s/big \(.*\) dog/small \1 cat/" filename

Does this help?

Code:

echo "a big black dog" | sed -e 's/big \(.*\) dog/small \1 cat/'
a small black eared cat

The \1 means that the first pattern that matches (.*) (any word) is put in. Similarly, another (second) matching pattern can also be put in. As an example:

Code:

echo "a big black eared dog ran away" | sed -e 's/big \(.*\) dog \(.*\)/small \1 cat \2/'
a small black eared cat ran away

Torch,

this is the not the way it works

Code:

echo "a big black dog" | sed -e 's/big \(.*\) dog/small \1 cat/'
a small black cat

Shouldn't the echo statement be as,

Code:

echo "a big black eared dog"

Ok guys, thanks for the confusion lol. Hope someone can clarify for sure.

Sed -e "s/big \(.*\) dog/small \1 cat/" filename

Explanations:

1> -e - is used when using multiple replacements usinf regular expressions or not
2> you are searching for "big" and "anything after that" and then "dog"
3> you are replacing "big" with "small" and "dog" with "cat" without deterioriating the "anything after that" text

when sed searches, it remembers the \(.*\) and it can be accessed using \1.

hope this helps.


Cheers,
Devaraj Takhellambam

Thanks Devtakh, that helped alot actually.

Sorry for the confusion. I copied the wrong output there.