Yannick Delahousse 

Based on exactly what you have shown here, without any test file to play with, and with no examples of what you are seeing and what you'd expect to see I can only hazard a guess.

You seem to be trying to exclude TUs based on a rule you created so that the target element would not be extracted for that TU.  But without commenting on the effectiveness of your rules for the TU (which are also ineffective) you are still extracting the target element for each time the target element is present.

You probably only need one rule for the whole file... something like this:

//trans-unit[not(contains(@id, 'metatag'))]/target

This should only extract the target if the trans-unit it belongs to doesn't have a metatag in the id attribute.  But as I said... completely untested.

Hi Paul
Thanks for your quick reply. 

Please see the file attached here with the filter. 

/cfs-file/__key/communityserver-discussions-components-files/90/JobID8626_5F00_en_5F00_zh_2D00_hans.zip

You are correct, I don't want the target element to be extracted for the TUs showing 'meta' and 'link' as they don't need translation.
I've asked the Webteam to exclude it from the XLF files but no success so far.
Thanks for your help 

Yannick Delahousse 

Thanks for the file... I note you didn't try to implement my suggestion.  So I just deleted all your rules and add just these two:

So this as non-translatable to make sure we don't take anything at all from the file:

//*

Then this to take just what we need:

//trans-unit[not(contains(@id, 'metatag'))]/target

Results in this which looks more like what you're after... unchecked though!

Hi Paul

I had tried your suggestions but now realize I forgot to move it up above my other rules for it to work. 

Your rule has gotten rid of the metatag content which is great improvement. But it's brought back other content I don't want to display in Trados like //note (for ex <note>Card Stack</note> seg21).

I also want to exclude center, image and bundle_default; Overview Text, Shield image; etc
I tried to copy your rule and use //trans-unit[not(contains(@id, 'rh_action'))]/target and //trans-unit[not(contains(@id, 'rh_redirect_fallback_action'))]/target but no success

Thanks for your help with this Paul

 Yannick Delahousse 

Yannick Delahousse said:

had tried your suggestions but now realize I forgot to move it up above my other rules for it to work. 

You should delete your other rules completely... I don't believe they are necessary and may introduce things you don't want.

Yannick Delahousse said:

But it's brought back other content I don't want to display in Trados like //note (for ex <note>Card Stack</note> seg21).

Not for me... segment #21 is taking "Card Stack" from the target element as required:

Yannick Delahousse said:

I also want to exclude center, image and bundle_default; Overview Text, Shield image; etc

Isn't this text written into the source and target element itself as translatable text? For example:

<target xml:lang="zh-hans"><![CDATA[center]]></target>

<target xml:lang="zh-hans"><![CDATA[image]]></target>

<target xml:lang="zh-hans"><![CDATA[bundle_default]]></target>

You need a completely different approach for this... which might be easier to just do with regex using the display filter!!

But if you really want a parser rule for it maybe something like this?

//trans-unit[not(contains(@id, 'metatag'))]/target[not(contains(., 'center') or contains(., 'image') or contains(., 'bundle_default'))]

I just added three... I'm sure you can add the rest.  But note this is still only two rules.  There is absolutely no need for all the other rules you have in there.  They serve no purpose at all other than to risk polluting what you really need.

 Yannick Delahousse 

Yannick Delahousse said:

had tried your suggestions but now realize I forgot to move it up above my other rules for it to work. 

You should delete your other rules completely... I don't believe they are necessary and may introduce things you don't want.

Yannick Delahousse said:

But it's brought back other content I don't want to display in Trados like //note (for ex <note>Card Stack</note> seg21).

Not for me... segment #21 is taking "Card Stack" from the target element as required:

Yannick Delahousse said:

I also want to exclude center, image and bundle_default; Overview Text, Shield image; etc

Isn't this text written into the source and target element itself as translatable text? For example:

<target xml:lang="zh-hans"><![CDATA[center]]></target>

<target xml:lang="zh-hans"><![CDATA[image]]></target>

<target xml:lang="zh-hans"><![CDATA[bundle_default]]></target>

You need a completely different approach for this... which might be easier to just do with regex using the display filter!!

But if you really want a parser rule for it maybe something like this?

//trans-unit[not(contains(@id, 'metatag'))]/target[not(contains(., 'center') or contains(., 'image') or contains(., 'bundle_default'))]

I just added three... I'm sure you can add the rest.  But note this is still only two rules.  There is absolutely no need for all the other rules you have in there.  They serve no purpose at all other than to risk polluting what you really need.

Hi Paul
Thanks a lot for your assistance here. I removed all the other rules now and will keep things as they are. Everything is working perfectly well now and is much tidier :-)

Excellent... thanks for confirming Yannick Delahousse