Sql Server Selecting Distinct records and OrderBy NEWID() - sql

Question

this is the stored procdure, I need to select Distinct Records and display them in a random order but I am facing an error that selecting Distinct can not be used with newid(), so how can I walk around this ?
USE [OtlobODR]
GO
/****** Object: StoredProcedure [OtlobFood].[ListOffersItems] Script Date: 11/18/2012 13:01:34 ******/
SET ANSI_NULLS ON
GO
SET QUOTED_IDENTIFIER ON
GO
ALTER procedure [OtlobFood].[ListOffersItems]
#Fk_CampaignID int
as
select distinct
CampaignID, CampaignOffers.OldPrice
, dbo.ItemSizes.IS_Id,
, dbo.Items.[Item_Description_L2] as Item_Description
, dbo.Items.[Item_Image]
, dbo.Items.[Item_Details]
, dbo.ItemSizes.[IS_Price] as Price
-- if null then it is not a featured meal
, dbo.ProviderItems.[PI_Id] as ProviderItems_PI_ID
, dbo.ItemCategories.[ItemCat_Id]
, dbo.Providers.Provider_Name_L2 as Provider_Name
, dbo.Providers.Provider_Menu_Logo
, dbo.Providers.Provider_Id
FROM Items
INNER JOIN ProviderItems ON Items.Item_Id = ProviderItems.Item_Id
INNER JOIN dbo.ItemSizes ON dbo.Items.Item_Id = dbo.ItemSizes.Item_Id
INNER JOIN CampaignOffers ON CampaignOffers.ItemID = ItemSizes.IS_Id
INNER JOIN dbo.ItemCategories ON dbo.Items.ItemCat_Id = dbo.ItemCategories.ItemCat_Id
INNER JOIN dbo.Providers ON dbo.ProviderItems.Provider_Id = dbo.Providers.Provider_Id
INNER JOIN dbo.Branches ON dbo.Providers.Provider_Id = dbo.Branches.Provider_Id
where Fk_CampaignID=#Fk_CampaignID
group by NEWID(),
CampaignID, CampaignOffers.OldPrice ,
dbo.ItemSizes.IS_Id,
dbo.Items.[Item_Description_L2],
dbo.Items.[Item_Image],
dbo.Items.[Item_Details],
dbo.ItemSizes.IS_Id,
dbo.ItemSizes.[IS_Price] ,
-- if null then it is not a featured meal
dbo.ProviderItems.[PI_Id] ,
dbo.ItemCategories.[ItemCat_Id],
dbo.Providers.Provider_Name_L2,
dbo.Providers.Provider_Menu_Logo
,dbo.Branches.Branch_Id,
dbo.Providers.Provider_Id,CampaignID,CampaignOffers.OldPrice
order by NEWID()

Answer 1

You need to push the SELECT DISTINCT into an inner query (at which point you can also lose the GROUP BY) then do a NEWID() in the outer query. The general form is
select
newid(), X.*
from
(
select distinct <cols>
from <tables>
where <whatever>
) X
order by 1
in your case I think this is what you want:
select
newid(), X.*
from
(
select distinct
CampaignID, CampaignOffers.OldPrice ,
dbo.ItemSizes.IS_Id,
dbo.Items.[Item_Description_L2] as Item_Description ,
dbo.Items.[Item_Image],
dbo.Items.[Item_Details],
dbo.ItemSizes.[IS_Price] as Price,
-- if null then it is not a featured meal
dbo.ProviderItems.[PI_Id] as ProviderItems_PI_ID,
dbo.ItemCategories.[ItemCat_Id],
dbo.Providers.Provider_Name_L2 as Provider_Name,
dbo.Providers.Provider_Menu_Logo,
dbo.Providers.Provider_Id
FROM
Items
INNER JOIN ProviderItems ON Items.Item_Id = ProviderItems.Item_Id
INNER JOIN dbo.ItemSizes ON dbo.Items.Item_Id = dbo.ItemSizes.Item_Id
inner join CampaignOffers ON CampaignOffers.ItemID = ItemSizes.IS_Id
INNER JOIN dbo.ItemCategories ON dbo.Items.ItemCat_Id = dbo.ItemCategories.ItemCat_Id
INNER JOIN dbo.Providers ON dbo.ProviderItems.Provider_Id = dbo.Providers.Provider_Id
INNER JOIN dbo.Branches ON dbo.Providers.Provider_Id = dbo.Branches.Provider_Id
where
Fk_CampaignID = #Fk_CampaignID
) X
order by 1

Answer 2

Remove the NEWID()'s from your query and surround it with
SELECT * FROM (
<your query>
) as t
ORDER BY NEWID()

Answer 3

This query doesn't make much sense, there is no aggregation, so what is the purpose of the group by statement? and grouping by a NEWID() is not going to get you any grouping's anyway.
Combine the superfluous group by with the distinct clause and it feels more like you have a join condition incorrect that is causing a cartesian, and you are desperately trying to use distinct and group by's to eliminate that cartesian, it certainly makes little sense to have them both in the statement, and the newid in the group by should be removed.