Eugene Elutin (7/25/2013)

---

webrunner (7/25/2013)

---

Thanks, Eugene, I see. I had commented out the ROLLBACK for my INSTEAD OF trigger so didn't understand the behavior you just posted with the test code....

Easy enough to demonstrate that:

USE tempdb --you can use a "real" db if you prefer

GO

IF EXISTS(SELECT 1 FROM sys.tables WHERE name = 'table1' AND schema_id = 1)

...

Eugene Elutin (7/25/2013)

---

Alexander Suprun (7/24/2013)

---

It's so untrue...

RAISERROR has nothing to do with the transaction.

It's...

webrunner (7/24/2013)

---

Alexander Suprun (7/24/2013)

---

Remove ROLLBACK from the trigger and lower severity of RAISERROR to 10, in this case SSMS won't show any errors, rows will disappear from the UI (but...

You don't need to explicitly cast it in this case, because integer has a higher precedence than varchar, so SQL will implicitly force the varchar to integer.

You should check the...

From a performance standpoint, a single table can work just as well as partitioning in many cases as long as you have the correct clustered index. But if you've...

This is why cursors exist. Just use a cursor, unless you have an extraordinarily high volume of emails to send.

CREATE TABLE [dbo].[TestData](

[text] NULL, --??...

You'll get better performance from clustering on the date -- as the only or first key column -- if date/date range is specified in the queries.

I believe a...

You don't need a CTE. In this case, it would just add complexity for no real reason that I can see :-).

Sean Lange (7/22/2013)

---

Yes you can still do a wildcard search.

inner join #search...

Michael Valentine Jones (7/22/2013)The CASE expression will verify that the data string is in YYYY-MM-DD format, and is a valid date. It will return a 1 if the date...

Something like this should be very close at least, I think:

create proc pSearch

@searchparam varchar(400) -- values will be entered like this - seperated by comma: John, Mary, 23.10.1980

as

set @searchparam...

chrissorric (7/17/2013)Thanks for all the help!

This is what I finally came up with

SELECTp.OrderNum

FROM@Orders AS p

INNER JOIN@Results AS x ON x.LabNum = p.LabNum

GROUP BYp.OrderNum

HAVING SUM(Case when x.Flag = 'N'THEN 1...

If the first date is greater than the second date, you will get a negative value.

If you want to treat all date differences as positive, add ABS() function around the...

If you can, instead just truncate the old table, then re-insert only the kept rows.