create blank file

[rhnaeco]

i know this may seem like an obvious question, but i am trying to create a blank text file from the command line (so that it doen't open), because i want to send continuous data into it from a loop ( i.e the first command is >> file.dat.
at the moment i am using :
$ echo > file.dat

which leaves a blank line at the beginning once data arrives.
i have tried just using >> to create the file but it produces an error for some of my input files so it seems easier to just make a blank file (which has a different name depending on the input data file).
I have also tried touch, but again i want to overwrite if the name exists.
thanks in advance

 

[Chacalinc]

just:

> file.txt

or

touch file.txt

cheers

Chacal, Inc.

 

[rhnaeco]

i knew it'd be simple! thanks!

 

[feherke]

Hi

echo -n > file.txt      [gray]# man echo : [b]-n[/b] do not output the trailing newline[/gray]
cp /dev/null file.txt   [gray]# DOS habit ( copy null file.txt )[/gray]

Feherke.

http://rootshell.be/~feherke/

 

[p5wizard]

The ">file.txt" won't work in c-shell (invalid null command) but other suggestions will.

Also some unix's echo commands don't understand -n option (check your man page), in that case:

echo "\c" >file.txt

to suppress the trailing newline...


HTH,

p5wizard

 

[Trevoke]

'touch', being created for this specific purpose is, I think, the cleanest solution.

-Haben sie fosforos?
-No tiengo caballero, but I have un briquet.

 

[feherke]

Hi

'touch', being created for this specific purpose is, I think, the cleanest solution.

I think, it is not. See the thread starter post :

I have also tried touch, but again i want to overwrite if the name exists.

Feherke.

http://rootshell.be/~feherke/

 

[SamBones]

This works for csh...

:> file.dat

It also works in other shells too. The ":" is the "null command".

Hope this helps.

 

[feherke]

Hi

A question : why to create an empty file, then append to it from a loop ? If your code looks like this :

for i in `seq 10`; do
  echo "$i*$i=$((i*i))" >> file.dat
done

Try to change it like this :

for i in `seq 10`; do
  echo "$i*$i=$((i*i))"
done > file.dat

The second is faster and, at least for me, easyer to maintain.

Feherke.

http://rootshell.be/~feherke/

 

[p5wizard]

- or -

logfile=/tmp/log.txt

rm -f ${logfile}

any command >>${logfile}

any other command >>${logfile}

while condition_cmd
do
 loop command
done >>${logfile}

The first >> redirection will create the file anew since it was deleted in the beginning of the script.

the rm -f will not complain if the file isn't there, will just silently do nothing in that case.

Note that using fixed names for logfiles basically makes your program single-user/single-task...


HTH,

p5wizard

 

[PHV]

One main reason to empty the file instead of recreating it is the preservation of original permissions/ownerships.

Hope This Helps, PH.
Want to get great answers to your Tek-Tips questions? Have a look at FAQ219-2884 or FAQ181-2886

 

[rhnaeco]

Feherke, i need to start with a blank file as it is called file$sample.dat, and at least for the time being, something is going horribly wrong with some of the maths in the loop and i need to see the files step by step.
also, i am still very new to scripting, and didn't think of that as a solution for when the maths DOES start working!

thanks to everyone for their solutions, this really is the best forum to come with a computer problem!

rhnaeco