2列のデータのすべての可能な組み合わせ

これが1つの方法です。

理論的には、製品ごとに最大20個のタグに対応できます（数値テーブルのサイズによって制限されます）が、私はそれをわざわざ試みませんでした。私のデスクトップでは、16個のタグを持つ単一の製品の65,535件の結果を作成するのに約30秒かかりました。うまくいけば、製品ごとの実際のタグ数はそれよりはるかに少なくなります！

IF OBJECT_ID('tempdb..#Nums') IS NULL
BEGIN
CREATE TABLE #Nums
(
i int primary key
)

;WITH 
L0 AS (SELECT 1 AS c UNION ALL SELECT 1),       
L1 AS (SELECT 1 AS c FROM L0 A CROSS JOIN L0 B), 
L2 AS (SELECT 1 AS c FROM L1 A CROSS JOIN L1 B), 
L3 AS (SELECT 1 AS c FROM L2 A CROSS JOIN L2 B),
L4 AS (SELECT 1 AS c FROM L3 A CROSS JOIN L3 B), 
L5 AS (SELECT 1 AS c FROM L4 A CROSS JOIN L4 B), 
Nums AS (SELECT ROW_NUMBER() OVER (ORDER BY (SELECT 0)) AS i FROM L5)

INSERT INTO #Nums
SELECT TOP 1048576 i FROM Nums;
END


;with ProductTags As
(
SELECT 1 ProductId,'Leather' AS Tag UNION ALL
SELECT 1, 'Watch' UNION ALL
SELECT 2, 'Red' UNION ALL
SELECT 2, 'Necklace' UNION ALL
SELECT 2, 'Pearl'
), NumberedTags AS
(
SELECT 
      ProductId,Tag,
       ROW_NUMBER() OVER (PARTITION BY ProductId ORDER BY Tag) rn,
       COUNT(*) OVER (PARTITION BY ProductId) cn
FROM ProductTags
),
GroupedTags As
(
SELECT ProductId,Tag,i
FROM NumberedTags
JOIN #Nums  on 
               #Nums.i <  POWER ( 2 ,cn)  
           and #Nums.i & POWER ( 2 ,rn-1) > 0
)
SELECT ProductId, 
       STUFF((SELECT CAST(', ' + Tag AS VARCHAR(MAX)) 
          FROM GroupedTags g2
         WHERE g1.ProductId = g2.ProductId and g1.i = g2.i
      ORDER BY Tag
       FOR XML PATH ('')),1,1,'') AS Tags
FROM GroupedTags g1
GROUP BY ProductId, i
ORDER BY ProductId, i

返品

ProductId   Tags
----------- ------------------------------
1            Leather
1            Watch
1            Leather, Watch
2            Necklace
2            Pearl
2            Necklace, Pearl
2            Red
2            Necklace, Red
2            Pearl, Red
2            Necklace, Pearl, Red