Shell Programming and Scripting

hi all, I have a question for you all. I have a xml file that contains 2 similar tags, like the example below,
<some tag>
<Tag ABC=1>
<tag_value1>some_value</tag_value1>
</Tag>
</some tag>
...
<Tag ABC=1>
<tag_value1>some_value</tag_value1>
</Tag>

My question would be, how do you retrieve the second <Tag ABC=1>, and then update the some_value right between the <tag_value1>?

I had created following scripts (a section of the whole scripts)
--------------------------------------------------------------------------
cat (file.xml) | while read data
newValue="<tag_value1>some_other_value</tag_value1>"
xmlHelper="echo $data"
echo "$xmlHelper" > $tempXML
strHelper="<Tag ABC=1>\n <tag_value1>some_value</tag_value1>"
oldValue=$(awk -F 'BEGIN{FS="<tag_value1>|</tag_value1>"}
/myMatch/{
for(i=1;i<=NF;i++) {
if [[ ${xmlHelper} = ${strHelper} ]]
then
print "<tag_value1>\(.*)</tag_value1>"
}
}
END {print $2}' file.xml)

if [[ ${xmlHelper} = ${strHelper} ]]; then
sed -e 's/$oldValue/newValue' file.xml
--------------------------------------------------------------------------

I am using ksh to do all these, the problem is that it would not run. Kept giving me an error of awk at line 2. Tried to go around but it come back the same way.

Wonder if anyone have a good suggestion?

Tou can try something like this (not tested) :

Code:

awk '
   /<TAG ABC=1>/ { 
       tagnum += 1 
   }
   /<tag_value1>/ && tagnum==2 { 
       print "<tag_value1>some_other_value</tag_value1>" ; 
       next 
   }
   1
' file.xml

Jean-Pierre.

well i had tried your section and replaced it with mine it does not work though giving me an error of,

awk: syntax error near line 10
awk: bailing out near line 10

But I had took out the number "1" it works fine now, still have a couple of errors to fix but i think i will be able to manage it, thank you for your time, aigles.

The number 1 alone in the awk script is the same thing as

Code:

{ print $0 }

Jean-Pierre.

ah ok... thank you. But the version of this awk in the ksh that i am using does not support that though. What I had done is as follow
awk '
/<TAG ABC=1>/ {
tagnum += 1
}
/<tag_value1>/ && tagnum==2 {
print $0;
next
}
' file.xml

Thank you for your help though, it had pushed me toward the right direction.

if on Solaris, try the original aigles' code with '/usr/bin/nawk' OR '/usr/xpg4/bin/awk' instead of thein old 'awk'

Nope i am running this on SunOS Generic_117350-33 sun4u sparc SUNW,Sun-Fire

well i am still fixing aigles' code to suite my need, because his codes will return all the related value, but all i want is just the correspondent value

i.e.
<some tag>
<Tag ABC = 1>
<tag_value1>some_value</tag_value1>
</Tag>
<some tag>
...
<Tag ABC = 1>
<tag_value1>some_value</tag_value1>
</Tag>

his scripts will return each and every instances that match <tag_value1> and the second condition will always be true, not really what I wanted, but extremely close (thank you aigles). I just wanted to extract the second

<Tag ABC = 1>
<tag_value1>some_value</tag_value1>
</Tag>

and the change the "some_value" to a new value.