STRING_SPLIT vs DelimitedSplitN4K


I think it is clear which option I must go with here. It was a nice ride with Delimited


		if OBJECT_ID('tempdb..#words') IS NOT NULL
			drop table #words

		create table #words(word nvarchar(4000))

		insert into #words
		select N'ከለታት አንድ ቀን በአንዲት ከተማ'

		select * 
	     from #words
		 cross apply sqlteam.dbo.DelimitedSplitN4K(word,' ')

		select * 
	     from #words
		 cross apply STRING_SPLIT(word,' ')

I don't know how you even see the spaces in your source data :slight_smile:


Utf8 glasses, must have fashion wear for all sql devs



Still, with STRING_SPLIT we don't know the order in which the tokens appear in the string, right?

That is correct James (though order is not an issue for me). But the glaring issue is it is breaking down words on each character. I think this has to also do with collation , digging into this further