Tier Query Results MS SQL

we do not see what you see: table structure/data/etc.. please clarify:
what is the result now and what you wish to have?

if you can send some sample \example with some create table\populate this table with some data and your select -- it could help too

What you need to do is add another column e.g. StockistPercentage and use CASE

when Stockists >= 10% then 1
when Stockists < 10% then 2

and do the Order By StockistPercentage, Total Desc


if you need further help to calculate percentage then let me know

Thank you Both,

My tables that form the query look like this --

CREATE TABLE [dbo].[SupplierSurveyInput](
	[ID] [int] IDENTITY(1,1) NOT NULL,
	[SupplierSurveyID] [int] NULL,
	[BrandID] [int] NULL,
	[PartnerID] [int] NULL,
	[CYearStockist] [char](1) NULL,
	[CYearNStockistWhy] [int] NULL,
	[CountryID] [int] NULL,
	[Saleability] [decimal](5, 2) NULL,
	[Quality] [decimal](5, 2) NULL,
	[Availability] [decimal](5, 2) NULL,
	[Delivery] [decimal](5, 2) NULL,
	[Returns] [decimal](5, 2) NULL,
	[RetailPrice] [decimal](5, 2) NULL,
	[ServiceSupport] [decimal](5, 2) NULL,
	[Rep] [decimal](5, 2) NULL,
	[Accounts] [decimal](5, 2) NULL,
	[CreditControl] [decimal](5, 2) NULL,
	[NextYearTrade] [char](1) NULL,
	[NoTradeWhy] [int] NULL,
	[Completed] [char](1) NULL
) ON [PRIMARY]

GO

SET ANSI_PADDING OFF
GO

ALTER TABLE [dbo].[SupplierSurveyInput] ADD  DEFAULT ('N') FOR [CYearStockist]
GO

ALTER TABLE [dbo].[SupplierSurveyInput] ADD  DEFAULT ('N') FOR [NextYearTrade]
GO

ALTER TABLE [dbo].[SupplierSurveyInput] ADD  DEFAULT ('N') FOR [Completed]
GO

Select allOpen in new window

And

CREATE TABLE [dbo].[Brand](
	[ID] [int] IDENTITY(1,1) NOT NULL,
	[SupplierID] [int] NOT NULL,
	[Name] [nvarchar](50) NOT NULL,
	[Add1] [nvarchar](255) NULL,
	[Add2] [nvarchar](255) NULL,
	[City] [int] NULL,
	[County] [int] NULL,
	[Country] [int] NULL,
	[Postcode] [nvarchar](10) NULL,
	[SwitchBoard] [nvarchar](20) NULL,
	[Fax] [nvarchar](20) NULL,
	[URL] [nvarchar](250) NULL,
	[Logo] [nvarchar](250) NULL,
	[UK] [char](1) NULL,
	[Eire] [char](1) NULL,
	[TGIAccountMgr] [int] NULL,
	[StartDate] [datetime] NULL,
	[ExitDate] [datetime] NULL
) ON [PRIMARY]

GO

SET ANSI_PADDING OFF
GO

ALTER TABLE [dbo].[Brand] ADD  DEFAULT ('N') FOR [UK]
GO

ALTER TABLE [dbo].[Brand] ADD  DEFAULT ('N') FOR [Eire]
GO

Select allOpen in new window

I can query the table using the query above, and I can query the table like so -

SELECT COUNT([ID])/62 Completed
  FROM [dbo].[SupplierSurveyInput]
  where CYearStockist <> 'u'

Select allOpen in new window

To get the total number of people that participated in the survey.

I'm stuck as to how to marry these two together?

Thanks

Can you also provide some sample data for both files in excel table format

thanks

Hi RehanYousaf,

I think I'm a stage closer on this, using the query below, what's the best way of including your CASE statement?

The query below adds a column that returns the total number of survey participants...

select 
ROW_NUMBER() OVER(order by (((sum(Saleability)  + sum(Quality)  + sum(Availability) + sum(Delivery) + sum([Returns]) + sum(RetailPrice) + sum(ServiceSupport) + sum(Rep) + sum(Accounts) + sum(CreditControl) ) /10)  / count(*)) desc) as Position, 
brandid,
B.Name,
sum(Saleability) /count(*) as avgSaleability,
sum(Quality) /count(*) as avgQuality,
sum(Availability) /count(*) as avgAvailability,
sum(Delivery) /count(*) as avgDelivery,
sum(Returns) /count(*) as avgReturns,
sum(RetailPrice) /count(*) as avgRetailPrice,
sum(ServiceSupport) /count(*) as avgServiceSupport,
sum(Rep) /count(*) as avgRep,
sum(Accounts) /count(*) as avgAccounts,
sum(CreditControl) /count(*) as avgCreditControl,
(((sum(Saleability)  + sum(Quality)  + sum(Availability) + sum(Delivery) + sum([Returns]) + sum(RetailPrice) + sum(ServiceSupport) + sum(Rep) + sum(Accounts) + sum(CreditControl) ) /10)  / count(*)) Total,
count(*) Stockists,
sum(case when nextyeartrade = 'Y' then 1 else 0 end) as StockistsNY, 
C.Completed Completed
from [dbo].[SupplierSurveyInput] S
inner join dbo.Brand B
on S.BrandID = B.ID
Left join(SELECT [SupplierSurveyID], COUNT([ID])/62 Completed FROM [dbo].[SupplierSurveyInput] where CYearStockist <> 'u' group by [SupplierSurveyID]) C on C.SupplierSurveyID = S.SupplierSurveyID
where [CYearStockist] = 'Y'
group by b.name, brandid, S.SupplierSurveyID, C.Completed
order by Total desc

Select allOpen in new window

The data is slightly complicated, but I'll see what I can do :)

If you have to calculate Stockists %, how will you do it

based on that I can provide you with the code


Also,  why are you hard coding 62 in your code

Here's the main table data, for the brand table, please remove all but the ID and Name columns for the DDL, so --

SET ANSI_NULLS ON
GO

SET QUOTED_IDENTIFIER ON
GO

SET ANSI_PADDING ON
GO

CREATE TABLE [dbo].[Brand](
	[ID] [int] IDENTITY(1,1) NOT NULL,
	[Name] [nvarchar](50) NOT NULL
) ON [PRIMARY]

GO

SET ANSI_PADDING OFF
GO

Select allOpen in new window

And use the Brands spreadsheet attached ....


Thank you :)SupplierSurveySample-Data.xlsxBrand-Sample-Data.xlsx

Thanks for the data, just need a little more explanation in simple term that
what Stockists is?,  and
to calculate Stockists percentage from the data provided in excel how will I do it using simple maths?

Thanks RehanYousaf, ok to explain, we are a buying group for a number of retailers (stockists) each year we ask them to complete a survey on the brands that they stock, for the data returned the 'Completed' column is the total number of retailers that completed the survey, 'Stockists' is the number of retailers that stocked the brand, as they don't stock all brands.

As we don't want the figures to be offset by a brand that has a very a single very happy stockist and no other stockists. we have decided to move all brands where less (or equal) than 10% of the the total 'completed' are stockists to the bottom and then apply the normal order by.

So in Excel, the query would be = Sum('Stockists'/'Completed')*100, when i tried this in SQL, I got strange results....

Hope this helps.

62 was hard coded, as we know that to be the total brands this year, although if this could be queried as COUNT DISTINCT  brandid that would be good also....

Thank you so much.

G

Your calculations are fine, just need to change

COUNT([ID])/62 Completed
to
CONVERT(FLOAT, COUNT([ID]))/62 Completed

and case statement will be something like

CASE
            WHEN c.Completed >= 0.10 THEN 1
            ELSE 0
      END AS PercentageOrder

Thank you, :)

I think we are almost there,.. I had just put the following together --

select 
ROW_NUMBER() OVER(order by (((sum(Saleability)  + sum(Quality)  + sum(Availability) + sum(Delivery) + sum([Returns]) + sum(RetailPrice) + sum(ServiceSupport) + sum(Rep) + sum(Accounts) + sum(CreditControl) ) /10)  / count(*)) desc) as Position, 
brandid,
B.Name,
sum(Saleability) /count(*) as avgSaleability,
sum(Quality) /count(*) as avgQuality,
sum(Availability) /count(*) as avgAvailability,
sum(Delivery) /count(*) as avgDelivery,
sum(Returns) /count(*) as avgReturns,
sum(RetailPrice) /count(*) as avgRetailPrice,
sum(ServiceSupport) /count(*) as avgServiceSupport,
sum(Rep) /count(*) as avgRep,
sum(Accounts) /count(*) as avgAccounts,
sum(CreditControl) /count(*) as avgCreditControl,
(((sum(Saleability)  + sum(Quality)  + sum(Availability) + sum(Delivery) + sum([Returns]) + sum(RetailPrice) + sum(ServiceSupport) + sum(Rep) + sum(Accounts) + sum(CreditControl) ) /10)  / count(*)) Total,
count(*) Stockists,
sum(case when nextyeartrade = 'Y' then 1 else 0 end) as StockistsNY, 
C.Completed Completed,
CASE WHEN (count(*)*100)/C.Completed <= 10 then 1
Else 2 END as Tier
from [dbo].[SupplierSurveyInput] S
inner join dbo.Brand B
on S.BrandID = B.ID
Left join(SELECT [SupplierSurveyID], COUNT([ID])/62 Completed FROM [dbo].[SupplierSurveyInput] where CYearStockist <> 'u' group by [SupplierSurveyID]) C on C.SupplierSurveyID = S.SupplierSurveyID
where [CYearStockist] = 'Y'
group by b.name, brandid, S.SupplierSurveyID, C.Completed
order by Tier desc, Total desc

Select allOpen in new window

Which appears to work, but it still has the hard coded 62 in it, any suggestions on the best way to not have this hardcoded?

Again, thank you so much for all your help.

G

Anotehr way of doing this will be

;WITH cteSupplierSurveyInput AS (
	SELECT
		SupplierSurveyID,
		BrandID,
		SUM(Saleability) /COUNT(*) AS avgSaleability,
		SUM(Quality) /COUNT(*) AS avgQuality,
		SUM(Availability) /COUNT(*) AS avgAvailability,
		SUM(Delivery) /COUNT(*) AS avgDelivery,
		SUM(RETURNS) /COUNT(*) AS avgReturns,
		SUM(RetailPrice) /COUNT(*) AS avgRetailPrice,
		SUM(ServiceSupport) /COUNT(*) AS avgServiceSupport,
		SUM(Rep) /COUNT(*) AS avgRep,
		SUM(Accounts) /COUNT(*) AS avgAccounts,
		SUM(CreditControl) /COUNT(*) AS avgCreditControl,
		(((SUM(Saleability)  + SUM(Quality)  + SUM(Availability) + SUM(Delivery) + SUM([RETURNS]) + SUM(RetailPrice) + SUM(ServiceSupport) + SUM(Rep) + SUM(Accounts) + SUM(CreditControl) ) /10)  / COUNT(*)) Total,
		COUNT(*) Stockists,
		SUM(CASE WHEN nextyeartrade = 'Y' THEN 1 ELSE 0 END) AS StockistsNY
	FROM 
		SupplierSurveyInput S	
	GROUP BY 
		S.BrandID, 
		S.SupplierSurveyID
),
cteSupplierSurveyCountFiltered AS(
	SELECT 
		SupplierSurveyID,
		COUNT(*) AS SupplierSurveyCountFiltered
	FROM	
		SupplierSurveyInput	
	WHERE
		CYearStockist != 'U'
	GROUP BY
		SupplierSurveyID
),
cteSupplierSurveyCount AS (
SELECT 
	SupplierSurveyID,
	COUNT(*) AS SupplierSurveyCount
FROM	
	SupplierSurveyInput	
GROUP BY
	SupplierSurveyID
)
SELECT 
	ROW_NUMBER() OVER(ORDER BY s.Total) AS RowNumber,
	*,
	CONVERT(FLOAT,f.SupplierSurveyCountFiltered)/c.SupplierSurveyCount * 100 AS Percentage,
	case 
		when CONVERT(FLOAT,f.SupplierSurveyCountFiltered)/c.SupplierSurveyCount * 100 >= 10 then 1
		else 0
	end AS PercentageOrder
FROM 
	cteSupplierSurveyInput s
	LEFT JOIN cteSupplierSurveyCountFiltered f ON s.SupplierSurveyID = f.SupplierSurveyID
	LEFT JOIN cteSupplierSurveyCount c ON s.SupplierSurveyID = c.SupplierSurveyID
	LEFT JOIN Brand b ON s.BrandID = b.ID
ORDER BY 
	PercentageOrder,
	Total

Select allOpen in new window

Thank you,. but that gives me some strange results... One question,  the query below works great now, with the exception of the 'Position' column, which is no longer correct,..

select 
ROW_NUMBER() OVER(order by (((sum(Saleability)  + sum(Quality)  + sum(Availability) + sum(Delivery) + sum([Returns]) + sum(RetailPrice) + sum(ServiceSupport) + sum(Rep) + sum(Accounts) + sum(CreditControl) ) /10)  / count(*)) desc) as Position, 
brandid,
B.Name,
sum(Saleability) /count(*) as avgSaleability,
sum(Quality) /count(*) as avgQuality,
sum(Availability) /count(*) as avgAvailability,
sum(Delivery) /count(*) as avgDelivery,
sum(Returns) /count(*) as avgReturns,
sum(RetailPrice) /count(*) as avgRetailPrice,
sum(ServiceSupport) /count(*) as avgServiceSupport,
sum(Rep) /count(*) as avgRep,
sum(Accounts) /count(*) as avgAccounts,
sum(CreditControl) /count(*) as avgCreditControl,
(((sum(Saleability)  + sum(Quality)  + sum(Availability) + sum(Delivery) + sum([Returns]) + sum(RetailPrice) + sum(ServiceSupport) + sum(Rep) + sum(Accounts) + sum(CreditControl) ) /10)  / count(*))*20 Total,
count(*) Stockists,
sum(case when nextyeartrade = 'Y' then 1 else 0 end) as StockistsNY, 
C.Completed Completed,
CASE WHEN (count(*)*100)/C.Completed <= 10 then 1
Else 2 END as Tier
from [dbo].[SupplierSurveyInput] S
inner join dbo.Brand B
on S.BrandID = B.ID
Left join(SELECT [SupplierSurveyID], COUNT([ID])/COUNT(DISTINCT(BrandID)) Completed FROM [dbo].[SupplierSurveyInput] where CYearStockist <> 'u' group by [SupplierSurveyID]) C on C.SupplierSurveyID = S.SupplierSurveyID
where [CYearStockist] = 'Y' AND S.SupplierSurveyID = 1 AND CountryID in (3)
group by b.name, brandid, S.SupplierSurveyID, C.Completed
order by Tier desc, Total desc

Select allOpen in new window

Any thoughts? I know that

ROW_NUMBER() OVER(order by (((sum(Saleability)  + sum(Quality)  + sum(Availability) + sum(Delivery) + sum([Returns]) + sum(RetailPrice) + sum(ServiceSupport) + sum(Rep) + sum(Accounts) + sum(CreditControl) ) /10)  / count(*)) desc) as Position, 

Select allOpen in new window

is no longer correct as it doesn't take into consideration the tier element... just not sure how to correct this?

Thanks

G

Close, but not quite, this starts 1 at the first result with a tier of 1, which means that the last 19 have a position of the first 19...

Thanks

Thank you so much for your support, for your reference, the last query was just missing the DESC!

Thank you Thank you and Happy New Year

Sorry I missed one as well

use
CONVERT(FLOAT,(COUNT(*))*100)/C.Completed
instead of
(COUNT(*)*100)/C.Completed

in your script