🚀 go-pugleaf

Melman@uiuc.edu (Melvin Wilson) writes:
: Help!
: 
: I have a very simple (and for the more experienced, stupid) question.
: 
: From the command prompt on a NeXT how could I search a text file and remove  
: every instance of a particular expression in that text file?  I've tried but 
: cannot figure out the apropriate format of the command.
: 

the two ways that come immediately to mind are to use the Unix batch
editor, called "sed" and its -d option to remove multicharacter 
expressions.  See the man pages for sed (or any good Unix book).  

If all you want to do is to remove a particular character (such as the
infamous ^M s which come from DOS files, then it's probably even easier
to use the Unix command "tr" (for "translate characters", see the man pages
for tr) and it's -d option to eliminate the characters.  For example, 
if you have a file called testfile wiht ^M s in it

>>>>>>tesfile<<<<<<<<

The rain in spain falls mainly on the plain.^M

The rain in spain falls mainly down the drain.^M


then the command 

tr -d '\015' <testfile  >outfile  

will remove the ^M s (which have character code \015)

>>>>>>>outfile<<<<<<<<<<<

The rain in spain falls mainly on the plain.

The rain in spain falls mainly down the drain.

Again, see the man pages for whichever is appropriate or drop me
a note containing the specific example. (Personal opinion:  "sed"
is like a microwave, i.e. it's one of those things you can get 
through life easily without, but once you have learned about it your
life is much better).  

Hope this is Helpful 

tom

In article <BsFuBt.D8I@news.cso.uiuc.edu> Melman@uiuc.edu (Melvin Wilson)  
writes:
> Hello,
> 
> Help!
> 
> I have a very simple (and for the more experienced, stupid) question.
> 
> From the command prompt on a NeXT how could I search a text file and remove  
> every instance of a particular expression in that text file?  I've tried but  
I  
> cannot figure out the apropriate format of the command.

type
	<inputfile >outputfile sed 's/old/new/g'
replacing inputfile and outputfile with their respective values.
The expression "old" will be replaced by "new". If old or new contain
any / characters, use a different character in the sed expression instead, like
	sed 's:old:new:g'

--
Volker Herminghaus-Shirai (vhs@darkcube.radig.de)

Computer industry: Industry in which the number of units sold of any
given product is inversely proportional to its technical excellence.
See also: MS-DOS, Windows, IBM-PC, X, QWERTY, 80x86.

In article <1992Aug6.102953.8775@darkcube.radig.de> vhs@darkcube.radig.de writes:
>In article <BsFuBt.D8I@news.cso.uiuc.edu> Melman@uiuc.edu (Melvin Wilson)  
>writes:
>> Hello,
>> 
>> Help!
>> 
>> I have a very simple (and for the more experienced, stupid) question.
>> 
>> From the command prompt on a NeXT how could I search a text file and remove  
>> every instance of a particular expression in that text file?  I've tried but  
>I  
>> cannot figure out the apropriate format of the command.
>
>type
>	<inputfile >outputfile sed 's/old/new/g'
>replacing inputfile and outputfile with their respective values.
>The expression "old" will be replaced by "new". If old or new contain
>any / characters, use a different character in the sed expression instead, like
>	sed 's:old:new:g'
>
>--
>Volker Herminghaus-Shirai (vhs@darkcube.radig.de)
>
>Computer industry: Industry in which the number of units sold of any
>given product is inversely proportional to its technical excellence.
>See also: MS-DOS, Windows, IBM-PC, X, QWERTY, 80x86.

The easy way to REMOVE things using sed is to replace them by a null.

This is accomplished by saying:

'cat oldfile | sed s/searchstring//g > outputfile'

This will QUICKLY remove every instance of your searchstring by replacing it
with NOTHING =)

Have fun!

Mark Neidengard
mneideng@cosmos.aca.calpoly.edu