Shell Scripting problems (newbie)

yawgmoth81 · 02-07-2003, 04:51 AM

I want to compare the first row of two files, take out the differences and merge the tow files together...

This is what I have:

cut -f1 $1 > temp1
cut -f1 $2 > temp2
diff -D ver temp1 temp2 > verschil.txt

this is the output file:

51185682482
51143082482
51185243482
52285682482
51235682482
#ifndef ver <- needs to be deleted
51185684562 <- needs to be deleted
#endif /* ver */ <- needs to be deleted
51185121482
#ifdef ver <- needs to be deleted
52286121482 <- needs to be deleted
#endif /* ver */ <- needs to be deleted

Is there a way in shell scripting to delete the lines "#ifndef ver" and "#endif /* ver */" and the number in between?

If so, what can I do?

TIA

vladkrack · 02-07-2003, 07:14 AM

Yes, there are several ways of doing that. A simpler solution is to use grep -v , like:

# grep -v "#ifdef ver" temp1

Hope that helps ...

yawgmoth81 · 02-07-2003, 07:19 AM

Could you be more specific?

I need to delete every occurence of these 3 lines in the file...

#ifdef ver
a number
#endif /* ver */

Can this be done?? And if so, how?

acid_kewpie · 02-07-2003, 07:27 AM

i'm lost.. is that the first row or the first column??

vladkrack · 02-07-2003, 07:32 AM

Grep will not show the lines that has the specified Regular Expression (that is inside "). Your comand will be something like:

# cut -f1 $1 | grep -v "#ifdef ver" | grep -v "52286121482" |grep -v "#endif /* ver */"> temp1

yawgmoth81 · 02-07-2003, 07:49 AM

But what if the number is different every time...??

As is the case in this file:

#ifndef ver
51185684562
#endif /* ver */
51185121482
#ifdef ver
52286121482
#endif /* ver */

I need to delete the "#ifdef ver", "#ifndef ver", "#endif /* ver */" and the numbers between them... In this case it's 51185684562 and 52286121482

vladkrack · 02-07-2003, 10:11 AM

Sorry, I didn't see it before. In that case will be easier to use sed:

# cat verschil.txt | tr "\n" § | sed 's/#ifdef ver§[0-9]*[0-9]§#endif \/\* ver \*\/§//g' | tr § "\n" > verschill.txt.new

Hko · 02-07-2003, 10:58 AM

This a little simpler, and maybe a bit more reliable:

Code:

mv verschil.txt verschil.bak
sed '/^#ifn\{0,1\}def ver/,/^#endif/d' verschil.bak >verschil.txt

yawgmoth81 · 02-10-2003, 05:30 AM

Thanks everybody, problem solved!!

HKO's option worked perfectly...

I've tested it on a bigger file and came to the conclusion that there also is an #else /* ver */

So i edited the code a bit and it works like a charm:

Code:

sed '/^#ifn\{0,1\}def ver/,/^#else/d' verschil.bak > verschil.txt
sed '/^#endif/d' verschil.txt > verschil.dat

Now it deletes from if to else and it deletes the single line endif

If someone has a better way of doing this, please tell me...

If not, then I think this topic can be closed, because everything works fine...

Thanks again everyone!

PS: I was trying something but it didn't work so well... Can I delete everything before #if(n)def and everything after #endif /* ver */ leaving everything inside these lines intact? I also need to delete the #ifndef, #else and #endif lines when I do that...

51185683111
51185683112
#ifndef ver
51185683113
#else /* ver */
52285683113
#endif /* ver */
51185683114
51185683115

Hko · 02-10-2003, 03:23 PM

Quote:

Originally posted by yawgmoth81
[...] Can I delete everything before #if(n)def and everything after #endif /* ver */ leaving everything inside these lines intact? I also need to delete the #ifndef, #else and #endif lines when I do that...

You mean the other way around?

Code:

sed -n '/^#ifn\{0,1\}def ver/,/^#endif/p' -e '/^#else/d' verschil.bak > verschil.txt

yawgmoth81 · 02-11-2003, 05:56 AM

Hmmm that doesn't work for me...

I'm stuck with this:

51185683111
51185683112
#ifndef ver
51185683113
#else /* ver */
52285683113
#endif /* ver */
51185683114
51185683115
#ifndef ver

I only want to have the numbers between #if and #else and I want to delete the rest, it works almost perfectly, there's only one small thing I can't delete the lines before the first #if.

As a small test I made this so I could see all the steps this script took:

Code:

sed '/^#endif/d' tijdens.txt > tijdens2.txt
sed '/^#else/,/^#ifn\{0,1\}def ver/d' tijdens2.txt > tijdens3.txt
sed '/^#ifn\{0,1\}def ver/d' tijdens3.txt > na5.txt

As you can see it first deletes all occurences of #endif, then it deletes everything between #else and #if, and after that it deletes the remaining #if. (usually the first one in the file)

In the example above my file will look like this:

51185683111
51185683112
51185683113

But I want it to look like this:

51185683113

Can this be done?

yawgmoth81 · 02-24-2003, 02:31 AM

For those of you who want to know:

I have a script doing precisely what I want it to do...

I solved my last problem (see post above this one) by inserting a "#begin" line in the beginning of the file that had to be processed.

That way I could even delete the first lines BEFORE the first "#ifn..." line.

I'm now trying to convert this script into AWK, if I have any problems with that I'll be sure to post it here...

Thanks for all the help...