Nice question!

Thanks!

😀

I don't like qotd too long.

But, BOL states that with SERIALIZABLE isolation level:

Other transactions cannot insert new rows with key values that would fall in the range of keys read...

sp_rename leave unaltered the definition, (see sys.syscomments)

create procedure My_proc

as

select 1 as a

go

SELECT * FROM sys.syscomments

WHERE object_name(id) = 'My_proc'

go

exec sp_rename My_proc,Myproc

SELECT * FROM sys.syscomments

WHERE object_name(id) = 'Myproc'

After the rename the definition...

+1

Wonderfull!!!

It, also, works with subquery:

SELECT * FROM (

SELECT MAX(object_id) AS ID,MAX(name) AS NAME FROM sys.objects

) AS o

But in this case it states a wrong type for column NAME(int,null)

I noticed it for columns, tables and functions, but I happily surprised for star!

Good!

😀

A little variation:

-- this runs

DECLARE @i int = 1, @j-2 int = 2;

SET @i += 2;

SELECT @j-2 * @i

GO

-- this raises error

DECLARE @i int = 1, @j-2 int = @i...

Hugo, you always are the NUMBER ONE!

😀

WOW, another qotd with a special offer with "SUMMER POINTS"!

😀

I read question and answers many times because of high score.

I was searching for the swindle!

But it's only a straightforward basic question.

😀

type_desc is Latin1_General_CI_AS_KS_WS

So, also, in a "case sensitive" database type_desc LIKE '%cons%' returns data.

I'd like also a SPLIT/JOIN function as in VB. So, given a tab/cr limited string the function returns a table and viceversa.

PURGECHAR removes a list of chars from a string.

Thank you for question!

Theoretically, any query would return different value!

In practice, Only ORDER BY NEWID() works for us.

grzegorz.mozejko (6/16/2014)

---

grzegorz.mozejko (6/16/2014)

---

very usefull, thx

a small correction, when name of first index column is begining part of second index column (add comma):

...

Replace:

STUFF((SELECT ', ' + COLUMN_DATA_KEY_COLS.name

with:

STUFF((SELECT ', ' + COLUMN_DATA_KEY_COLS.name + ' ' +SUBSTRING('+-',INDEX_COLUMN_DATA_KEY_COLS.is_descending_key+1,1)

Thanks for the Article.

I voted it with 5 stars.

But there's only a hole:

Your query doesn't consider the order on columns.

Example:

create index idx_MyTab on MyTab(col1,col2)

and

create index idx_MyTab_1 on MyTab(col1 DESC,col2)

So, in...