grep option

[George2]

Hello guys,


I am new to this tool and I read some manual of it. I am wondering what are your favorite options to grep or frequently used? Could you share some to me -- a newbie of this tool? :-)

Just want to learn some best practices.


thanks in advance,
George

[Wraithan]

You could have just appended this to your previous thread on Grep. Just an observation.

[Mario F.]

I don't think I have "favorite options" for grep. I use it very seldom. Last time I did was maybe 4 months ago when I decided to rename a variable name I used on several source files.

In essence, this is mostly how grep is used. It's not a tool you use too often (or your shouldn't use too often), or that you can say you have a favorite expression.

[George2]

Thanks Mario,


From some manual I found, grep -r option is useful when we assign a directory parameter.

Originally Posted by Mario F.

I don't think I have "favorite options" for grep. I use it very seldom. Last time I did was maybe 4 months ago when I decided to rename a variable name I used on several source files.

In essence, this is mostly how grep is used. It's not a tool you use too often (or your shouldn't use too often), or that you can say you have a favorite expression.

regards,
George

[CornedBee]

"Favourite option?

You use what's necessary. If grep without options doesn't do what you need, read the manual and see if there is something.

[matsp]

Some quite useful things are:
"grep -ri something . " - search for "something" in this and all subdirectories, ignoring case.
"grep something file|grep -v someother" - search for something, but only on lines that also has someother in them.
"grep something file|grep -v someother" - search for something, but only on lines that DOESN'T have someother.
"grep ^something file(s)" - search for lines that start with something.

There are millions of other variations, but particularly chaining grep and grep -v can be quite handy for narrowing down a large search to a small one. E.g. you know there is a #define called "something", but you can't find which header file it is in, and "something" is used quite frequently in your code, so use "grep -r something .|grep #define" would show you lines that have something and is a #define.

You can of course write expressions that do this using regular expression syntax, but I find it easier to just chain together many grep's.

I sometimes use the "-5" or some such to see lines AROUND what I find.

But there is no such thing as "my favourite grep command". That's like saying "which size screwdriver is your favourite". Unless it's for burglary purposes [bigger is better here!], it's the one that has the right profile(Philips, Pozi, flat, Torx, etc) and size to fit the screw. Don't matter if I think the Pozi #1 is nicer looking, it won't help me screw in a Pozi #3 screw any better than "grep -v" will find what you were looking for in a file [in fact, the latter will show all lines that DON'T match].

--
Mats

[George2]

Thanks Mats,


Great experience sharing!

In the following sample, why not using grep -r "#define something" . directly?

Originally Posted by matsp

There are millions of other variations, but particularly chaining grep and grep -v can be quite handy for narrowing down a large search to a small one. E.g. you know there is a #define called "something", but you can't find which header file it is in, and "something" is used quite frequently in your code, so use "grep -r something .|grep #define" would show you lines that have something and is a #define.

--
Mats

regards,
George

[matsp]

Originally Posted by George2

Thanks Mats,


Great experience sharing!

In the following sample, why not using grep -r "#define something" . directly?




regards,
George

Because for example, you don't know the number of spaces (or if it's a tab), or when it's #define something someother, and you are searching for "someother" may be a good example.

Typically, you get "error -5", and you want to know which NAME error 5 is, so you search for -5 and then the #define.

--
Mats

[George2]

Thanks Mats,

Originally Posted by matsp

Because for example, you don't know the number of spaces (or if it's a tab), or when it's #define something someother, and you are searching for "someother" may be a good example.

Typically, you get "error -5", and you want to know which NAME error 5 is, so you search for -5 and then the #define.

--
Mats

My question is answered.


regards,
George

[esbo]

I don't know about grep but I got quite good at regular expressions, for example to reverse the order of two columns of words in a file it was something like

:%s/{*[0-9a-Z]} {*[0-9a-Z]}/\\2 \\1/

Or something like that.
That is almost certaintly wrong though as I have not used regular expressions for ages.
But I could do a lot of things then which are rather time consuming to do in a 'modern' editor.

I used to use then to make my programs meet the required 'standard' of programming style or
'convention'. So I would have to change all my variable and function names etc..

[George2]

Thanks esbo,


What do you think is the best regular expression tutorial for a newbie, so that I can grasp and understand what you wrote below in a day (at least understand the same level)? :-)

Originally Posted by esbo

I don't know about grep but I got quite good at regular expressions, for example to reverse the order of two columns of words in a file it was something like

:%s/{*[0-9a-Z]} {*[0-9a-Z]}/\\2 \\1/

Or something like that.
That is almost certaintly wrong though as I have not used regular expressions for ages.
But I could do a lot of things then which are rather time consuming to do in a 'modern' editor.

I used to use then to make my programs meet the required 'standard' of programming style or
'convention'. So I would have to change all my variable and function names etc..

regards,
George

[esbo]

Originally Posted by George2

Thanks esbo,


What do you think is the best regular expression tutorial for a newbie, so that I can grasp and understand what you wrote below in a day (at least understand the same level)? :-)




regards,
George

Not realy sure George it is really a long time since I used them, hence the example I gave
almost certaintly contains errors.

To change cat to dog would be

s/cat/dog/

to di more than once on a line

s/cat/dog/g

on every line

%s/cat/dog/g

to capture cat and put it after dog

%s/\{cat\}/dog \1/g

Not sure if the above is correct and the syntax might vary in applications soI guess you wil have to try google

This looks OK
http://www.robelle.com/smugbook/regexpr.html

My examples were for search and relace in an editor. (vi)
The equivilent command in unix is 'sed' I believe.

[George2]

Thanks esbo!


Looks great! Your below comments, "di" means?

Originally Posted by esbo

to di more than once on a line

s/cat/dog/g

regards,
George

[esbo]

di = do

Typing error, 'i' is next to 'o'!!

[George2]

Cool, thanks esbo!

Originally Posted by esbo

di = do

Typing error, 'i' is next to 'o'!!

regards,
George