SQL UPDATE (but only if its going to be unique) - sql

Question

I need to write an SQL string that will UPDATE a value into a table, but only if that value is unique. But it should allow blank values ('') to be duplicates.
UPDATE [knw].[dbo].[Nop_ProductVariant]
SET barcode = (CASE WHEN (SELECT count(*) FROM [knw].[dbo].[Nop_ProductVariant] WHERE barcode = '" + item + "') = 0 THEN '" + item + "' ELSE '' END)
WHERE ProductVariantId='" + s + "'"
This functions seems to wipe all data and make them blank (''). I think this is because if the current value that I am changing is the same value, (i.e. changing 'purple' to 'purple') it thinks it would be a duplicate and puts a blank one in.
Help!

Answer 1

Could you modify the SELECT inside your CASE slightly to exclude the current ProductVariantId?
SELECT count(*)
FROM [knw].[dbo].[Nop_ProductVariant]
WHERE barcode = '" + item + "'
AND ProductVariantId <> '" + s + "'"

Answer 2

I guess you want something like this:
UPDATE [knw].[dbo].[Nop_ProductVariant]
SET barcode = #NewValue
WHERE NOT EXISTS
(SELECT 1
FROM [knw].[dbo].[Nop_ProductVariant]
WHERE barcode = #NewValue)
AND ... ? ;

Answer 3

Why not simply run it in two queries:
Query 1:
Update [knw].[dbo].[Nop_ProductVariant]
Set barcode = '"' + [knw].[dbo].[Nop_ProductVariant].item + '"'
Where ProductVariantId='"' + s + '"'
And Exists (
Select 1
From [knw].[dbo].[Nop_ProductVariant] As PV1
Where PV1.barcode = '"' + [knw].[dbo].[Nop_ProductVariant].[item] + '"'
)
Query 2:
Update [knw].[dbo].[Nop_ProductVariant]
Set barcode = ''
Where ProductVariantId='"' + s + '"'
And Not Exists (
Select 1
From [knw].[dbo].[Nop_ProductVariant] As PV1
Where PV1.barcode = '"' + [knw].[dbo].[Nop_ProductVariant].[item] + '"'
)

Answer 4

Normally, you should use a UNIQUE index for such situations.
In SQL Server 2008:
CREATE UNIQUE INDEX ux_productvariant_barcode ON (barcode) WHERE barcode <> ''
If you are not able to create such an index, use this:
WITH q AS
(
SELECT *
FROM [knw].[dbo].[Nop_ProductVariant]
WHERE ProductVariantId = #s
)
UPDATE q
SET barcode = #item
WHERE #item NOT IN
(
SELECT barcode
FROM q
WHERE barcode <> ''
)

Answer 5

This shows all unique non-empty barcodes along with their first-entry IDs:
SELECT barcode, MIN(ProductVariantId) AS ProductVariantId
FROM [knw].[dbo].[Nop_ProductVariant]
WHERE barcode <> ''
GROUP BY barcode
Accordingly, this shows all their duplicates (full rows this time, not including the IDs from the previous selection):
SELECT npv.*
FROM [knw].[dbo].[Nop_ProductVariant] AS npv
LEFT JOIN (
SELECT MIN(ProductVariantId) AS ProductVariantId
FROM [knw].[dbo].[Nop_ProductVariant]
WHERE barcode <> ''
GROUP BY barcode
) npv_u ON npv.ProductVariantId = npv_u.ProductVariantId
WHERE npv.barcode <> '' AND npv_u.ProductVariantId IS NULL
Personally I would first check out those lists. If they are alright, the following script can be used to blank the duplicates:
UPDATE npv
SET barcode = ''
FROM [knw].[dbo].[Nop_ProductVariant] AS npv
LEFT JOIN (
SELECT MIN(ProductVariantId) AS ProductVariantId
FROM [knw].[dbo].[Nop_ProductVariant]
WHERE barcode <> ''
GROUP BY barcode
) npv_u ON npv.ProductVariantId = npv_u.ProductVariantId
WHERE npv.barcode <> '' AND npv_u.ProductVariantId IS NULL