Replace Text Based On Pattern Matching

[Grizzly521]

Hi Im trying to replace text in a file based upon a pattern.

The pattern Im looking for is:

<styleURL>#style0002</styleURL>
<name>#######6105#######</name>

The # are seven alphanumeric characters before and after 6105.

I need it to replace that with:

<styleURL>#style0003</styleURL>
<name>#######6105#######</name>

Any ideas would be appreciated.

Thanks

 

[feherke]

Hi

You not gave enough details, so I let you to finish it :

awk 'NR>1{if(p~/style0002/&&$0~/6105/)sub(/style0002/,"style0003",p);print p}{p=$0}END{print p}' /input/file

Tested with [tt]gawk[/tt] and [tt]mawk[/tt].

Feherke.

http://rootshell.be/~feherke/

 

[Grizzly521]

What more information do you need?

 

[feherke]

Hi

Well, there could be a few questions rose by this sentence :

The # are seven alphanumeric characters before and after 6105.

I am wondering why that was important.
[ul]
[li]Is the # in front of style0002 also a digit ?[/li]
[li]Are there other style0002 occurrences not enclosed in styleURL tags ?[/li]
[li]Are there other style0002 occurrences not followed by 6105 enclosed in name tags ?[/li]
[li]Are the style0002 and the 6105 always on separate lines ?[/li]
[li]Are the style0002 and the 6105 always on consecutive lines ?[/li]
[li]Is there a formula to find out the new text like newtext=oldtext+1 in your example ?[/li]
[/ul]


Feherke.

http://rootshell.be/~feherke/

 

[Grizzly521]

Thanks for the pointers. I didnt realize how vague I had been.

* Is the # in front of style0002 also a digit ?
No. I overlooked that when I picked # to represent variables
* Are there other style0002 occurrences not enclosed in styleURL tags ?
Yes.
* Are there other style0002 occurrences not followed by 6105 enclosed in name tags ?
No.
* Are the style0002 and the 6105 always on separate lines ?
Yes.
* Are the style0002 and the 6105 always on consecutive lines ?
In the format I showed yes.
* Is there a formula to find out the new text like newtext=oldtext+1 in your example ?
No.

What Im looking to replace always looks like

<styleURL>#style0002</styleURL>
<name>*******6105*******</name>

and needs to be turned into

<styleURL>#style0003</styleURL>
<name>*******6105*******</name>

The * are alphanumeric characters, seven before and after 6105.

 

[feherke]

Hi

Well, then you only have to complete my previous code to restrict the circumstances in which the replacement is performed :

awk [highlight]--re-interval[/highlight] 'NR>1{if(p~/[highlight]<styleURL>#[/highlight]style0002[highlight]<\/styleURL>[/highlight]/&&$0~/[highlight]<name>[[:digit:]]{7}[/highlight]6105[highlight][[:digit:]]{7}<\/name>[/highlight]/)sub(/[highlight]<styleURL>#[/highlight]style0002[highlight]<\/styleURL>[/highlight]/,"[highlight]<styleURL>#[/highlight]style0003[highlight]</styleURL>[/highlight]",p);print p}{p=$0}END{print p}' /input/file

Some notes :
[ul]
[li]The --re-interval option is [tt]gawk[/tt] only.[/li]
[li]Some [tt]awk[/tt] implementations not handle POSIX character classes like [tt][:digit:][/tt].[/li]
[li]Some [tt]awk[/tt] implementations not handle [tt]{n}[/tt] quantifiers.[/li]
[/ul]
So another question would be, what kind of [tt]awk[/tt] implementation you have ? Anyway, this is a more portable alternative :

awk 'NR>1{if(p~/<styleURL>#style0002<\/styleURL>/&&$0~/<name>[highlight][0-9][0-9][0-9][0-9][0-9][0-9][0-9][/highlight]6105[highlight][0-9][0-9][0-9][0-9][0-9][0-9][0-9][/highlight]<\/name>/)sub(/<styleURL>#style0002<\/styleURL>/,"<styleURL>#style0003</styleURL>",p);print p}{p=$0}END{print p}' /input/file

Tested with [tt]gawk[/tt] and [tt]mawk[/tt].

If neither that works, you could change some syntax elements and you get the [tt]perl[/tt] code :

perl -ne 'if($.>1){if($p=~m/<styleURL>#style0002<\/styleURL>/&&$_=~m/<name>[[:digit:]]{7}6105[[:digit:]]{7}<\/name>/){$p=~s/<styleURL>#style0002<\/styleURL>/<styleURL>#style0003<\/styleURL>/}print$p}$p=$_;END{print$p}' /input/file

Feherke.

http://rootshell.be/~feherke/

 

[Grizzly521]

Ill give it a try at work tonight. I really appreciate it.

Itd be running on Solaris 10.
I cant remember what version Im running of the top of my head.

 

[feherke]

Hi

If you will hit [tt]awk[/tt] problems on Solaris, this advice usually helps :

Try using nawk or /usr/xpg4/bin/awk instead.

Feherke.

http://rootshell.be/~feherke/