Comments on: SQL SERVER – Introduction to Function SIGN

By: tejas

tejas — Mon, 11 Feb 2013 07:53:33 +0000

SUM( DECODE( SIGN(DAYDIFF(TransactionReceivableDetail.Receivable.AssignedPayment.CreatedOn,TempAttributesRescission.RescissionPaymentAsofDate)),1,0,TransactionReceivableDetail.Receivable.AssignedPayment.PaymentAmount))

this query is returning a null value can u explain

By: Eric Low

Eric Low — Sat, 08 Dec 2012 17:43:37 +0000

Hi! This is SQL statement found in SQL Profiler..

SELECT SUM(amount_book * -SIGN(cleared_type -0.5)) FROM cminpdtl WHERE date_applied <= 734760 AND cash_acct_code = '15080202ADMNHQIP002'

May I know what is the use of -SIGN(fieldname -0.5)?
Why must use -0.5 and not just the fieldname will do? cleared_type is just a value either 0 or 1.

Thank you very much in advance!

By: pinaldave

pinaldave — Thu, 28 Jun 2012 11:14:49 +0000

Looks excellent!

By: Spyder

Spyder — Thu, 28 Jun 2012 10:25:30 +0000

Brilliant.

I have to run a report where all values must be positive. Now, instead of:

UPDATE MyTable
SET Amount =
CASE
WHEN Amount < 0 THEN
(Amount * -1) * Currency
ELSE
Amount * Currency
END

I can use:

UPDATE MyTable
SET Amount = (Amount * SIGN(Amount)) * Currency

Seems to run quicker too. Any comment on that?

By: pinaldave

pinaldave — Wed, 27 Jun 2012 07:11:45 +0000

Excellent suggestion – I totally agree with it!

By: pinaldave

pinaldave — Wed, 27 Jun 2012 07:11:19 +0000

Hi,

In that case you can just multiply the value with -1.

By: guillegr123

guillegr123 — Tue, 26 Jun 2012 14:14:24 +0000

There is a kind of “inverse” function of the SIGN function? For example, if the value is a positive number, it returns -1, and if it is a negative number, it returns 1.

By: pinaldave

pinaldave — Tue, 26 Jun 2012 05:16:00 +0000

I re-ran the query and new image is now uploaded. Thanks for your attention to the details.

By: vincyf1

vincyf1 — Tue, 26 Jun 2012 05:07:13 +0000

My I suggest another correction: NumVal3 value in the screenshot should be 0.00 :)

By: vincyf1

vincyf1 — Tue, 26 Jun 2012 05:05:39 +0000

Result for NumVal3 should be 0.00 in the screenshot :)

By: Kumar Harsh

Kumar Harsh — Mon, 25 Jun 2012 05:10:58 +0000

Sign can also be use in
i)When we have to find if table contain any rowcount or not.
Select sign(@@rowcount)
Select sign(count(*) from emp
ii) IsNullOrEmpty(equivalent function of c#)

CREATE FUNCTION dbo.IsNullOrEmpty(@text NVARCHAR(4000))
RETURNS BIT
AS
BEGIN

return SIGN(LEN(isnull(@text,”)))

END

By: pinaldave

pinaldave — Mon, 25 Jun 2012 03:04:09 +0000

Fixed the typo thank you so much My friend!

By: pinaldave

pinaldave — Mon, 25 Jun 2012 03:03:52 +0000

Fixed the typo thank you so much!

By: kk

kk — Mon, 25 Jun 2012 01:11:50 +0000

Thanks for writing about this function. I have learnt a new function. But seems you have a typo in the second SELECT statement:
SIGN(@NumVal2) NumVal3 => SIGN(@NumVal3) NumVal3

By: Michael Emerick

Michael Emerick — Sun, 24 Jun 2012 16:41:55 +0000

Think you might have a small typo, though the results you have portrayed are correct.

This:
SELECT SIGN(@NumVal1) NumVal1,SIGN(@NumVal2) NumVal2,SIGN(@NumVal2) NumVal3

Should be:
SELECT SIGN(@NumVal1) NumVal1,SIGN(@NumVal2) NumVal2,SIGN(@NumVal3) NumVal3

As always thanks so much for all of your posts, they have helped me come a long way!

By: DHall

DHall — Sat, 23 Jun 2012 05:19:01 +0000

Sign() can also be used it calculate an equivalent alternative to the ABS() function. Note that the last two columns have the same values.:

USE tempdb
CREATE TABLE T ( N INT )
INSERT INTO T VALUES ( -42), (-11), (0), (NULL), (17), (2)
SELECT N,
N * SIGN(N) AS NxSignN,
ABS(N) AS AbsN
FROM T
DROP TABLE T
GO

By: jay

jay — Sat, 23 Jun 2012 02:57:11 +0000

Excellent post. It’s impressive how SIGN allows one to write much less code to get the same output.