今天在项目中遇到的这个问题，确实纠结我挺久的。其实这种需求如果用代码来实现的话只需要一个典然后就能实现，然而，有时候就偏偏让你写个sql，然后就去执行下就OK了，因此，问题就来了(标题挺乱的，也取得很不好)
其实这里的思想很简单，就是分列插入临时表，然后处理临时表，最终得到结果
语文能力很差，那就直接上代码：
/*******原需求：
--1 导出所有四级（省、市、区/县、镇/街道）、所属区号、工作时长、能否取货、能否派货、所属分部
--1.1 若“所属区号、上班时间、下班时间、是否有记录、所属分部”存在不同的值，则并列显示（逗号隔开）
********/
--第一步：取出所需字段，并去重
select distinct DR_250,DR_251,DR_252,DR_260,DR_234,DR_254,DR_273,DR_274,DR_208
into #TMP from TB_DriverDomainNew
--第二步：取单个差异字段，并保存到临时表中
SELECT distinct DR_250,DR_251,DR_252,DR_260,DR_208 INTO #TMP1 FROM #TMP--选择有DR_208不相同的记录
SELECT distinct DR_250,DR_251,DR_252,DR_260,DR_234 INTO #TMP2 FROM #TMP--选择有DR_234不相同的记录
SELECT distinct DR_250,DR_251,DR_252,DR_260,DR_254 INTO #TMP3 FROM #TMP--选择有DR_254不相同的记录
SELECT distinct DR_250,DR_251,DR_252,DR_260,DR_273 INTO #TMP4 FROM #TMP--选择有DR_273不相同的记录
SELECT distinct DR_250,DR_251,DR_252,DR_260,DR_274 INTO #TMP5 FROM #TMP--选择有DR_274不相同的记录
--第三步：合并不停记录，并用逗号隔开
select DR_250,DR_251,DR_252,DR_260,
(select ',' + DR_208 from #TMP1 as A where B.DR_260 = A.DR_260 AND B.DR_251 = A.DR_251 AND B.DR_250 = A.DR_250 AND B.DR_252 = A.DR_252 for XML path('')) AS DR_208
INTO #TMP1_1 from #TMP1 as B
GROUP BY DR_250,DR_251,DR_252,DR_260
select DR_250,DR_251,DR_252,DR_260,
(select ',' + DR_234 from #TMP2 as A where B.DR_260 = A.DR_260 AND B.DR_251 = A.DR_251 AND B.DR_250 = A.DR_250 AND B.DR_252 = A.DR_252 for XML path('')) AS DR_234
INTO #TMP2_2 from #TMP2 as B
GROUP BY DR_250,DR_251,DR_252,DR_260
select DR_250,DR_251,DR_252,DR_260, --数值类型需要转换convert(nvarchar(3),DR_254)
(select ',' + convert(nvarchar(3),DR_254) from #TMP3 as A where B.DR_260 = A.DR_260 AND B.DR_251 = A.DR_251 AND B.DR_250 = A.DR_250 AND B.DR_252 = A.DR_252 for XML path('')) AS DR_254
INTO #TMP3_3 from #TMP3 as B
GROUP BY DR_250,DR_251,DR_252,DR_260
select DR_250,DR_251,DR_252,DR_260,
(select ',' + DR_273 from #TMP4 as A where B.DR_260 = A.DR_260 AND B.DR_251 = A.DR_251 AND B.DR_250 = A.DR_250 AND B.DR_252 = A.DR_252 for XML path('')) AS DR_273
INTO #TMP4_4 from #TMP4 as B
GROUP BY DR_250,DR_251,DR_252,DR_260
select DR_250,DR_251,DR_252,DR_260,
(select ',' + DR_274 from #TMP5 as A where B.DR_260 = A.DR_260 AND B.DR_251 = A.DR_251 AND B.DR_250 = A.DR_250 AND B.DR_252 = A.DR_252 for XML path('')) AS DR_274
INTO #TMP5_5 from #TMP5 as B
GROUP BY DR_250,DR_251,DR_252,DR_260
--第四步：连接获取所有记录
SELECT T1.DR_250,T1.DR_251,T1.DR_252,T1.DR_260,T1.DR_208,T2.DR_234,t3.DR_254,T4.DR_273,T5.DR_274 INTO #TMP_F FROM #TMP1_1 T1
LEFT JOIN #TMP2_2 T2 ON T2.DR_260 = T1.DR_260 AND T2.DR_251 = T1.DR_251 AND T2.DR_250 = T1.DR_250 AND T2.DR_252 = T1.DR_252
LEFT JOIN #TMP3_3 T3 ON T3.DR_260 = T1.DR_260 AND T3.DR_251 = T1.DR_251 AND T3.DR_250 = T1.DR_250 AND T3.DR_252 = T1.DR_252
LEFT JOIN #TMP4_4 T4 ON T4.DR_260 = T1.DR_260 AND T4.DR_251 = T1.DR_251 AND T4.DR_250 = T1.DR_250 AND T4.DR_252 = T1.DR_252
LEFT JOIN #TMP5_5 T5 ON T5.DR_260 = T1.DR_260 AND T5.DR_251 = T1.DR_251 AND T5.DR_250 = T1.DR_250 AND T5.DR_252 = T1.DR_252
--第五步：去掉开头逗号，得到最终结果
SELECT DR_250,DR_251,DR_252,DR_260
,SUBSTRING(DR_234, 2, Len(DR_234)-1) DR_234
,SUBSTRING(DR_254, 2, Len(DR_254)-1) AS DR_254
,SUBSTRING(DR_273, 2, Len(DR_273)-1) AS DR_273
,SUBSTRING(DR_274, 2, Len(DR_274)-1) AS DR_274
,SUBSTRING(DR_208, 2, Len(DR_208)-1) AS DR_208
FROM #TMP_F
WHERE ISNULL(DR_260, '')<>''
--第六步：删临时表
DROP TABLE #TMP
DROP TABLE #TMP1
DROP TABLE #TMP2
DROP TABLE #TMP3
DROP TABLE #TMP4
DROP TABLE #TMP5
DROP TABLE #TMP_F
DROP TABLE #TMP1_1
DROP TABLE #TMP2_2
DROP TABLE #TMP3_3
DROP TABLE #TMP4_4
DROP TABLE #TMP5_5
代码看起来确实非常凌乱，这确实是我本身的原因，但是还是希望对大家有所帮助，也是我自己的一个知识积累，希望下次遇到的时候能够更快的实现