Hi Pietro,

Not strange at all! Try using this regex in the filter: [A-Za-z]

OMG !!!
IT WORKS AMAZINGLY!

THANK YOU NORA!
THANK YOU PAUL!!!

Guys ...
I still need your help.
Actually there is something that is not really working or I am missing something ...
[A-Za-z] filter is not enough ... would You have a look please?

THANK YOU one more time!
Pietro

PIETRO SFERRINO

Time to get the text books out!

Try this and see if it works better:

^(?=[A-Za-z]+)

Would have been much better if you provided some sample text in a file and not just a screenshot.

Well ... Paul ... Actually, I am starting to ...

By the way ...
Sure; here is a Sample Document.

DOCX

Paul said:

^(?=[A-Za-z]+)

Well ... This works MUCH better ...

I'll keep trying.
Thank You

Post Scriptum ...
By the way ... I really need to get out of a Virtual Machine !!!
It works terrible!

May I ask you 3 things ???

1.
Is trados a 32 bit program ONLY ??? (whandering why I ahve a 32 bit version installed)
(This come from the Activity Manager, once it often crashes)

2.
Do You know how Studio 2019 works on ARM processors ??? If it is not optimized then I will not take a comp working on that processor ...

3.
Just to give an example ...



SORRY!
And ... THANK YOU AGAIN!
Pietro

PIETRO SFERRINO

You could also try an approach like this for this particular file:

(?=[А-Яа-я]+|^[\d\.+-]+)

Then reverse the result:

Just to give you some other ideas of how to use this filter.  It's quite easy to build up the filter based on the results you see in the Editor until you get just what you need.

Or even this and reverse it if you wanted to copy source to target... not sure if that all works for you but it might help you see how I used it:

(?=[А-Яа-я]+)

PIETRO SFERRINO said:

1.
Is trados a 32 bit program ONLY ??? (whandering why I ahve a 32 bit version installed)
(This come from the Activity Manager, once it often crashes)

Yes.  There is little to be gained in moving to 64-bit until some more fundamental improvements (that the teams are working on) have been implemented.

PIETRO SFERRINO said:

2.
Do You know how Studio 2019 works on ARM processors ??? If it is not optimized then I will not take a comp working on that processor ...

I have no idea.  Maybe David Watson could provide a sensible answer to that question?

And I don't really know what the third question is?

PIETRO SFERRINO

You could also try an approach like this for this particular file:

(?=[А-Яа-я]+|^[\d\.+-]+)

Then reverse the result:

Just to give you some other ideas of how to use this filter.  It's quite easy to build up the filter based on the results you see in the Editor until you get just what you need.

Or even this and reverse it if you wanted to copy source to target... not sure if that all works for you but it might help you see how I used it:

(?=[А-Яа-я]+)

PIETRO SFERRINO said:

1.
Is trados a 32 bit program ONLY ??? (whandering why I ahve a 32 bit version installed)
(This come from the Activity Manager, once it often crashes)

Yes.  There is little to be gained in moving to 64-bit until some more fundamental improvements (that the teams are working on) have been implemented.

PIETRO SFERRINO said:

2.
Do You know how Studio 2019 works on ARM processors ??? If it is not optimized then I will not take a comp working on that processor ...

I have no idea.  Maybe David Watson could provide a sensible answer to that question?

And I don't really know what the third question is?

\p{IsCyrillic} can also be used as range.

Hello Anthony!
Yes. This works too.
Thank You!

Or obviously the reverse:  \P{IsCyrillic} -> not Cyrillic

Yes ... Thanks
I am trying to find a way to match only that segments having at the same time both cyrilli AND non-cyrillic; or ascii AND non-ascii symbols ... BUT, I can not find a way ... have to trick more; bu I can see that RegEx take a while to be understood ... it is not something so easy. But still interesting!

AND conditions can often be handled with a series of Lookarounds. 

As a starter
(?=.*\P{IsCyrillic}+)(?=.*\p{IsCyrillic}+)
but this needs to be refined to handle bounds and things like spaces, punctuation.
What do you mean by "non-ASCII symbols"?

Anthony Rudd said:

What do you mean by "non-ASCII symbols"?

It's a strange thing to ask for but I'd assume anything that didn't fall within this table:

https://www.ascii-code.com/

If you ignore the extended characters then you're talking about the first 127 character codes.  So to recognise ASCII symbols you could use this:

[\x00-\x7F]

To recognise non-ASCII symbols you could use this:

[^\x00-\x7F]

If you want to include the extended character codes then you need to go all the way up to 255.  

In this case to recognise ASCII symbols you could use this:

[\x00-\xFF]

To recognise non-ASCII symbols you could use this:

[^\x00-\xFF]

PIETRO SFERRINO

I'd also encourage you to create separate threads for your questions and don't just keep tacking them on the end of this one.  Since your questions are different each time, and since we mark the verified answers so other users can easily find the solution to the original posted question, it would be far more helpful if you helped and didn't just create an ongoing list of Q&A with no mechanism for anyone to identify which question was actually being answered.

Just create them in here... if they are also regex or xpath related:

community.sdl.com/.../regex_and_xpath

Ok, Thank You Paul. I'll do so.
Yes, abotu Ascii and non ascii ... I already had done that ... simply some problem using the AND to find segments where both russian and non-russian words are ... I'll try later and if I'll need I'll open a new thread! Thanks.
Pietro

I am returning to this question, because it was not fully answered and the solution includes some important considerations.

Because the \P{IsCyrillic} block is too inclusive (it contains more than just alphabetic characters), the list of excluded non-Cyrillic punctuation characters would need to be adapted. A better approach is to define the set of word characters (all alphabetic characters + numbers + underscore) that excludes Cyrillic characters.

The following regex matches a segment that contains Cyrillic and non-Cyrillic characters (while excluding digits and various miscellaneous characters):

(?=.*[\w-[\p{IsCyrillic}\d]]+)(?=.*[\p{IsCyrillic}]+)

The regex matches the first line, but not the last two lines:

White automobile "белый автомобиль" in Russian‹┘

White automobile in Russian‹┘

"белый автомобиль"