Shell Programming and Scripting

Hi,

I'm trying to extract XML values for a particular field (MessageId) from a file that has multiple XML data on many lines. For example:

Line1: <?xml version = "1.0"?><xml></xml><xml></xml><xml></xml><xml></xml><xml></xml><field name='MessageId'><value>123456789</value></field><xml></xml><xml></xml><xml></xml><xml></xml>
Line2:<?xml version = "1.0"?><xml></xml><xml></xml><xml></xml><field name='MessageId'><value>987654321</value></field><xml></xml><xml></xml><xml></xml>
.
.
Line n

Please note: MessageId does not appear in the same position from line to line. The red text is the same format for each line.

I haven't been able to find anything on the forums for my situation. Everything I've tried with sed has been returning the whole line as opposed to just the string of numbers for MessageId. My goal is to store them in an array so I can do searches on another data file.

Any help would be greatly appreciated!

Thanks,
Matt

are you expecting this ?!

Code:

grep -o "<field name='MessageId'>.*</field>" t1
<field name='MessageId'><value>123456789</value></field>
<field name='MessageId'><value>987654321</value></field>

Kindly put the input in code tags !

It certainly is not standard.

The man page says "show only the part of the line that matches PATTERN"?

That's what SED is for!

I was at least hoping that the -o option wasn't greedy, but it still grabs what it can.

Is there a UUOO (useless use of options award)?

Yep, can't use grep -o.


Any other suggestions?

Thanks!

both sed /awk suffers from greediness. another approach

Code:

awk 'BEGIN{FS="</field>"}
{
   for (i=1;i<NF;i++){
    gsub(/.*fieldname=/,"",$i)
    print $i
   }
}' file