Sandor Juhasz 

It can be done:

The trick is seeing what the segments look like without the rule.  If you turn on the show spaces you'll see that the first space is not there because it's considered unnecessary and is moved external to the segment.  So you only really need a rule for the "dash+space".

To do that I usually put something in the basic editor so I know where the rule needs to be like this:

Then switch to Advanced and replace the XX and YY with a proper regular expression.  I know you probably don't need to do this, but I've done it for so long it makes more sense in my head!

So this:

XX[YY]+

Would be become this:

-[\s]+

So before the break is the dash, and break on the space.

Hi Paul,

Thanks for your reply.

This rule will separate everything after a space+dash:

"- Solar panels project - separate Project?" =>

"-"

"Solar panels project -"

"separate Project?"

I would need to separate the starting dash only:

"-"

"Solar panels project - separate Project?"

Is it possible?

Sandor Juhasz 

No problem... ish:

Although I did have to play around a little because Trados Studio can be a bit unpredictable with regex for segmentation.  I expected any of these to work but they failed:

^-[\s]+
(?<!.)-[\s]+

So in the end made it work with this:

^\s*-[\s]+

The \s* accounts for any potential quirks in Trados' handling of invisible characters (e.g., tabs, extra spaces), and anchoring with ^ ensures the match applies only to the first hyphen at the start of the line.

Unfortunately, this doesn't work for me.

My segmentation rule looks like this:

And the segmented file (I've also attached the source file):

2768.example.xlsx

 Sandor Juhasz 

Imagine how much time you would save us both by providing a sample in the first place!  Not anything I imagined from your description!!

I'm sorry but I'm not experienced with adding placeholder tags. What should I do?

Sandor Juhasz 

Like this:

1. enable embedded content processing in your excel filetype (before you create your project!)
2. select "extract in all paras"
3. add a new placeholder rule (\s-\s)
4. edit the rule
5. click on advanced
6. set the segmentation to exclude

Thanks a lot Paul. It works now!

By the way, why did the segmentation rule not work? Maybe because of the soft returns? (I also have a segmentation rule for manual line break set.)

Sandor Juhasz 

Sandor Juhasz said:

By the way, why did the segmentation rule not work? Maybe because of the soft returns? (I also have a segmentation rule for manual line break set.)

Yes, I figured that, so I added one as well to be able to reproduce your problem.  I think the use of all segmentation rules is probably possible, but I didn't spend any time trying to figure it out as I think the filetype solution is easier.  It might even be possible to solve your entire problem using embedded content and not touch the segmentation rules.  But yes, I think the soft returns (if that's what they really are) probably complicated the issue.  We sort of evolved the eventual solution from each step because we didn't have that file to start with.  Generally I prefer to find the simplest way to solve these things as they are so often one-offs.

Sandor Juhasz 

Sandor Juhasz said:

By the way, why did the segmentation rule not work? Maybe because of the soft returns? (I also have a segmentation rule for manual line break set.)

Yes, I figured that, so I added one as well to be able to reproduce your problem.  I think the use of all segmentation rules is probably possible, but I didn't spend any time trying to figure it out as I think the filetype solution is easier.  It might even be possible to solve your entire problem using embedded content and not touch the segmentation rules.  But yes, I think the soft returns (if that's what they really are) probably complicated the issue.  We sort of evolved the eventual solution from each step because we didn't have that file to start with.  Generally I prefer to find the simplest way to solve these things as they are so often one-offs.