Pick some letters from string to fill field username (SHP 2013 workflow)

use regular expression action and store result in a single line text variable:

here is the pattern as text:

^([A-Z][a-zA-Z]{1,2})[a-z]* ([A-Z][a-zA-Z]{1,4}).*$

and the explanation:

Thank you very much for your response and viable screens and manual.

Probably, I'll have to study about the scheme you sent me. I hope I will solve the problem out with variables.

Edit: It does not work for me right now.

Edit 2: I already got it (I mean the system how it works). But it needs some precise work. Thank you for your pattern mlauer, but now what? I've pasted the pattern in "Pattern zone" and it still does not work. It shoud all be set correctly. Do not I need to put some "fn-replace" or something in front of the pattern? I also used Nintex tutorials to find awesome "RegExr" editor. It helped me a lot.

Edit 3: Your solution is working, but only when I use specified text, in our case, "Steve Jonasson". But I need to put into the field, where Steve Jonasson is, variable.

Check the attachement to see more. I'm going to try to solve it out by myself with variables, but every piece of advice may help me :)

Thanks.

Steve

Your screenshot looks good. Please note that for a test with "Run now" no variables can be used, because they are only available when starting the workflow.

You can extend the character classes enclosed in [] by unicode characters, e.g.

[x{00C4}-x{00C9}] matches range between Ä (index 196) and É (index 201)

see also List of Unicode characters

We do not understand each other.

Our language has accented letters/characters = ěščřžýáíéúůŠČŘŽÝÁÍÉÚŮ

The pattern doesn't accept it. I'm trying to reach the state when "Sežer Hřbet" will be put in collection like: "Sezer Hrbet" with no accent. At the final state, with your pattern, stored in collection "HrbetSez", it would be great.

Does anything like this exist? Can I do it with regex (regular expression)?