unusual uniq results

unusual uniq results

am 28.09.2004 17:09:12 von William Stanard

In doing a demo before a class (Linux Red-hat 2.4.18-14), I used the uniq
command on a file (dup_nums) that consisted of twelve lines, each line
containing a number, from one to 9. I repeated the numbers, 6, 8, and 9.
The std output showed the expected list of numbers, all duplicates
removed. At a student's suggestion, I ran uniq again, but this time
directed the output to the file itself....

uniq dup_num > dup_num

....and, lo and behold, the file dup_num was empty. I checked uniq --help,
info uniq, and man uniq, all to no avail. Shouldn't uniq have removed the
duplicate numbers and written one through 9 to the file?

Bill stanard




-
To unsubscribe from this list: send the line "unsubscribe linux-newbie" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.linux-learn.org/faqs

Re: unusual uniq results

am 28.09.2004 18:11:30 von Stone

The redirection takes place before your command. So basically what
you told the system to do was:

1. Create the file dup_num. If the file exists overwrite it.
2. Redirect stdout of "uniq dup_num" into dup_num.

Since you'd already overwritten the data in dup_num, there was nothing
for uniq to process.

On Tue, 28 Sep 2004 11:09:12 -0400, William Stanard
wrote:
> In doing a demo before a class (Linux Red-hat 2.4.18-14), I used the uniq
> command on a file (dup_nums) that consisted of twelve lines, each line
> containing a number, from one to 9. I repeated the numbers, 6, 8, and 9.
> The std output showed the expected list of numbers, all duplicates
> removed. At a student's suggestion, I ran uniq again, but this time
> directed the output to the file itself....
>
> uniq dup_num > dup_num
>
> ...and, lo and behold, the file dup_num was empty. I checked uniq --help,
> info uniq, and man uniq, all to no avail. Shouldn't uniq have removed the
> duplicate numbers and written one through 9 to the file?
>
> Bill stanard
>
> -
> To unsubscribe from this list: send the line "unsubscribe linux-newbie" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.linux-learn.org/faqs
>



--
http://xstonedogx.heroesmarket.net
-
To unsubscribe from this list: send the line "unsubscribe linux-newbie" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.linux-learn.org/faqs

Re: unusual uniq results

am 29.09.2004 13:07:08 von William Stanard

I'm confused. To me the > has always meant to take what is to the left of
the sign and redirect its output to whatever is to the right. In my case
this would mean taking the output of uniq dup_num and redirecting it. Can
you set me straight?
-- Bill

Stone writes:
>The redirection takes place before your command. So basically what
>you told the system to do was:
>
>1. Create the file dup_num. If the file exists overwrite it.
>2. Redirect stdout of "uniq dup_num" into dup_num.
>
>Since you'd already overwritten the data in dup_num, there was nothing
>for uniq to process.
>
>On Tue, 28 Sep 2004 11:09:12 -0400, William Stanard
> wrote:
>> In doing a demo before a class (Linux Red-hat 2.4.18-14), I used the
>uniq
>> command on a file (dup_nums) that consisted of twelve lines, each line
>> containing a number, from one to 9. I repeated the numbers, 6, 8, and 9.
>> The std output showed the expected list of numbers, all duplicates
>> removed. At a student's suggestion, I ran uniq again, but this time
>> directed the output to the file itself....
>>
>> uniq dup_num > dup_num


Bill Stanard
Academic Computing
Palmer Trinity School
305.969.4239



-
To unsubscribe from this list: send the line "unsubscribe linux-newbie" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.linux-learn.org/faqs

Re: unusual uniq results

am 29.09.2004 14:44:34 von Simon Valiquette

William Stanard a =E9crit :
> I'm confused. To me the > has always meant to take what is to the lef=
t of
> the sign and redirect its output to whatever is to the right.

That's basically what it does. Actually, it also create an empty=20
file or erase it if it already exist.

> In my case
> this would mean taking the output of uniq dup_num and redirecting it.=
Can
> you set me straight?
> -- Bill
>

It's the opposite. It creates the redirection first, even before=20
knowing if you ran a valid command.

What you don't seems to understand, is that the ">" ask the shell to=
=20
create an empty file _before_ executing uniq. It means that the data i=
n=20
"dup_num" are already overwritten when uniq is executed, and thus uniq=20
is opening an empty file.


Simon Valiquette
http://gulus.USherbrooke.ca



> Stone writes:
>=20
>>The redirection takes place before your command. So basically what
>>you told the system to do was:
>>
>>1. Create the file dup_num. If the file exists overwrite it.
>>2. Redirect stdout of "uniq dup_num" into dup_num.
>>
>>Since you'd already overwritten the data in dup_num, there was nothin=
g
>>for uniq to process.
>>
>>On Tue, 28 Sep 2004 11:09:12 -0400, William Stanard
>> wrote:
>>
>>>In doing a demo before a class (Linux Red-hat 2.4.18-14), I used the
>>
>>uniq
>>
>>>command on a file (dup_nums) that consisted of twelve lines, each li=
ne
>>>containing a number, from one to 9. I repeated the numbers, 6, 8, an=
d 9.
>>>The std output showed the expected list of numbers, all duplicates
>>>removed. At a student's suggestion, I ran uniq again, but this time
>>>directed the output to the file itself....
>>>
>>>uniq dup_num > dup_num
>=20
-
To unsubscribe from this list: send the line "unsubscribe linux-newbie"=
in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.linux-learn.org/faqs

Re: unusual uniq results

am 29.09.2004 16:53:19 von Yapo Sebastien

Well ... the main problem is that the bash redirection takes effect before
uniq is executed.

When you do "uniq dup_num > dup_num" the following happens :

- the bash redirection operator ">" triggers the opening of dup_num in
write-only mode (O_WRONLY | O_CREAT | O_TRUNC) to redirect stdout to the
file, thus truncating dup_num to zero length as the file already exists
- the file dup_num is opened read-only by uniq (O_RDONLY)
- the output stream is assumed by uniq to be stdout (no "OUTFILE" specified)
- each line of dup_num is read and treated by uniq but dup_num is empty
(truncated to zero length before) so, obviously, the result of uniq is also
empty.

This is probably not perfectly accurate but is "grosso modo" what happens.
Quote from bash manual (Redirection / Redirecting Output) :
"If the file does not exist it is created; if it does exist it is truncated to
zero size."

Sebastien

On Wednesday 29 September 2004 11:07, you wrote:
> I'm confused. To me the > has always meant to take what is to the left of
> the sign and redirect its output to whatever is to the right. In my case
> this would mean taking the output of uniq dup_num and redirecting it. Can
> you set me straight?
> -- Bill
>
> Stone writes:
> >The redirection takes place before your command. So basically what
> >you told the system to do was:
> >
> >1. Create the file dup_num. If the file exists overwrite it.
> >2. Redirect stdout of "uniq dup_num" into dup_num.
> >
> >Since you'd already overwritten the data in dup_num, there was nothing
> >for uniq to process.
> >
> >On Tue, 28 Sep 2004 11:09:12 -0400, William Stanard
> >
> > wrote:
> >> In doing a demo before a class (Linux Red-hat 2.4.18-14), I used the
> >
> >uniq
> >
> >> command on a file (dup_nums) that consisted of twelve lines, each line
> >> containing a number, from one to 9. I repeated the numbers, 6, 8, and 9.
> >> The std output showed the expected list of numbers, all duplicates
> >> removed. At a student's suggestion, I ran uniq again, but this time
> >> directed the output to the file itself....
> >>
> >> uniq dup_num > dup_num
-
To unsubscribe from this list: send the line "unsubscribe linux-newbie" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.linux-learn.org/faqs