SQL Server CTEs are not bad recursion, because they are not recursion at all.

Admitted, they are written as recursion, but they are a loop.

To get this clear, consider the following:

step...

Think STUFF is not a very clean way to convert XML dataype to varchar, so,

why not use cast( ... as varchar(max)), or, consistent with xml datatype, xml_thing.value ?

1. I never use static cursor but cursor variable

declare @C cursor

set @C =...

this article is a rare example for what i expect of a well-written article.

exposure at the beginning, conclusion, references, some reasonable reasoning in the main part.

congratulations.

---------------------

...........................

10 ^38 > 8 ^21 = 2 ^63

What about using 2 factors,

one for conversion from source into days and the other for conversion from days to target ?

1 more multiply but shorter program

PLEASE, could you, just in one sentence, tell the essence of what you are doing ?

I use this approach :

MERGE statement and comparing columns with SET ANSI NULLS OFF.

Any objections against using SET ANSI NULLS OFF ?

Will a total scan of an index use anything but the lef level ?

Just understood:

a nonclustered index may be the same size but never larger than a clustered index, so, if it covers the data the nonclustered index will be first choice.

was not sure about the outcome, tried it out :

all five of them used the nonclustered index (SQl Server 2012).

This is a little bit of a surprise, but also...

it's

TRY_CONVERT ( data_type [ ( length ) ], expression [, style ] )

liked this one.

used most of the features since, should have used all but one.

by the way: is there a forum or platform keeping track of desirable new features in TSQL,

ons...