Len function and nvarchar

Question

Hi we run 2012 enterprise and to be proactive i thought i'd ask the community what the len function will return on our nvarchar(30) col when our foreign partners start populating with chars that really need to occupy 2 bytes per visible char. 
I already see that len does an rtrim.   That's interesting but let's say they send a col value that occupies all 30 display positions with characters that require 2 bytes each?
Why do i ask?  First of all I see that our unicode data that contains chars that otherwise only require 1 byte count a fully populate column as len 30.  But I see max length in sql's meta data = 60.  And I will be applying the len function
to this column in some of our etl and need to anticipate correctly the behavior that i will see under any condition including a mix of chars that would otherwise occupy 1 byte together with chars that require two bytes in storage.
Depending on the answer i may need to know from the community what function I would use to count visible chars instead of used bytes?  Its the visible chars I really want to count.

Answer 1

As other have said: len() returns the number of characters and there is no difference between characters from Latin or Chinese scripts.
What I like to add is that is not that simple that varchar = 1 byte per character and nvarchar = 2 byte per character. In varchar characters from East Asian scripts typically take up two bytes per character; obviously you need the right collation for that.
And with nvarchar, if you use any of the new _SC collations that were added in SQL 2012, a character can take up two or four bytes. However an nvarchar(30) is not 30 characters - it's 60 bytes.
In case you use _SC collations, I should add that the characters that take up four bytes are fairly exotic. I believe that all living languages are well served by the base plane in Unicode.
Erland Sommarskog, SQL Server MVP, [email protected]