Grep for tab character

[Robert Mark Bram]

Hi All!

I cannot seem to use grep for tabs. Here is what I have tried..

This shows that there are tabs in the file test.txt:

Robert Mark Bram - /tmp
$cat test.txt | sed s/[\\t]/tab/g
tabdr
tabHow
tabHow
tabHow
tabHow
How
How

And these lines show how I am trying to grep for tab, but with no luck!

Robert Mark Bram - /tmp
$grep -e [\\t]How test.txt

Robert Mark Bram - /tmp
$grep -e [\t]How test.txt

Robert Mark Bram - /tmp
$grep -e "[\t]How" test.txt

Robert Mark Bram - /tmp
$grep -e "[\\t]How" test.txt

Robert Mark Bram - /tmp
$

Am I missig something? Or is it cygwin grep?

Any advice would be most appreciated!

Rob
:)
--
Robert Mark Bram
http://phd.netcomp.monash.edu.au/RobertMarkBram/default.asp
B.Comp.(Systems Development/Business Systems)
B.Net.Comp.(Hons)
Doctor of Philosophy Student

School of Network Computing
Faculty of Information Technology
Monash University
Peninsula Campus
McMahons Rd
Frankston, VIC 3199
AUSTRALIA

[Christopher Faylor]

Post by Robert Mark Bram
I cannot seem to use grep for tabs. Here is what I have tried..
And these lines show how I am trying to grep for tab, but with no luck!
Robert Mark Bram - /tmp
$grep -e [\\t]How test.txt
Robert Mark Bram - /tmp
$grep -e [\t]How test.txt
Robert Mark Bram - /tmp
$grep -e "[\t]How" test.txt
Robert Mark Bram - /tmp
$grep -e "[\\t]How" test.txt
Robert Mark Bram - /tmp
$
Am I missig something? Or is it cygwin grep?

How about just using the actual tab character? I don't see any indication that
grep is supposed to treat '\t' specially and it seems to behave that way on linux,
too.

What's wrong with just using tab?

cgf

[Robert Mark Bram]

Hi Christopher,

Thank you for your response!

Post by Christopher Faylor
How about just using the actual tab character? I don't see any indication that
grep is supposed to treat '\t' specially and it seems to behave that way on linux,
too.

I have read in many places that \t is a metacharacter for tab in regular
expressions - but maybe that's only for sed, perl, awk etc...
http://sitescooper.org/tao_regexps.html

Post by Christopher Faylor
What's wrong with just using tab?

Because that starts up auto-complete. In other words, I don't know how to
insert a literal tab into a command line.

But I do not like that idea for scripts either - it can be difficult to
visually discriminate a tab from a space character, which can too easily
lead to errors.

Rob
:)
--
Robert Mark Bram
http://phd.netcomp.monash.edu.au/RobertMarkBram/default.asp
B.Comp.(Systems Development/Business Systems)
B.Net.Comp.(Hons)
Doctor of Philosophy Student

School of Network Computing
Faculty of Information Technology
Monash University
Peninsula Campus
McMahons Rd
Frankston, VIC 3199
AUSTRALIA

[Brian Dessent]

Post by Robert Mark Bram

Post by Christopher Faylor
How about just using the actual tab character? I don't see any indication that
grep is supposed to treat '\t' specially and it seems to behave that way on linux,
too.

I have read in many places that \t is a metacharacter for tab in regular
expressions - but maybe that's only for sed, perl, awk etc...
http://sitescooper.org/tao_regexps.html

Try "grep -P '\t'" to use perl-compatible regexps. Note that this is a
specific capability of GNU grep, so it will not be portable to systems
that use a different grep. It might be more portable to use "awk
'/\t/'".

Or <ctrl-v><tab> to insert a literal tab as others said.

Brian

[fergus]

Try
grep "<Ctrl-v><Ctrl-i>"
i.e. type 4 characters being quotes <Ctrl-v> <Ctrl-i> quotes.
I think that works.
(I'm not certain that just typing <Tab> always (ever?) works, for me
anyway.)
Fergus

[Buchbinder, Barry (NIH/NIAID)]

Post by Robert Mark Bram
Because that starts up auto-complete. In other words, I don't know
how to insert a literal tab into a command line.

<control-V><tab>

Post by Robert Mark Bram
But I do not like that idea for scripts either - it can be difficult
to visually discriminate a tab from a space character, which can too
easily lead to errors.

Then try

sed -n '/\t/p' file