我从XML文件中检索了一组数据，现在我想将它们解析为不同的记录，以插入到我已经由不同列定义的表中。
在这个例子中，我发现源XML文件中有11个不同的列。我已经找到了列的相应排序顺序，以便它们始终以相同的顺序插入到表中。
不同的表格列示例
1 ad1_pk
2 ad1_address1
3 ad1_address2
4 ad1_address3
5 ad1_ctyfk
6 ad1_postalcode
7 ad1_active
8 ad1_irstat
9 ad1_irdata
10 ad1_at1fk
11 ad1_mk
使用以下代码，我已经准备好了数据来定义记录＆＃34;：
SELECT TOP 100 t.drn,
z.ID,
null RecNum,
z.Name,
z.Value
--INTO zTempXMLDataConverted2
FROM dbo.zTempXMLDataConverted z
JOIN #tempSortOrder t ON t.node = z.Name
ORDER BY z.id,t.drn
...但是，我只有一个问题：我可以直观地看到＆＃34;看到＆＃34;不同的记录，但我不知道如何以编程方式确定每条记录（也就是将＆＃34; RecNum＆＃34;列设置为适当的记录值。
所以，这最终是我想要看到的：
drn ------- ID ------ RecNum ------ Name -------------- Value
1 ------ 1 ------ 1 ------ ad1_pk --- 1
2 ------ 2 ------ 1 ------ ad1_address1 --- P.O. Box 5036
3 ------ 3 ------ 1 ------ ad1_address2 --- NULL
4 ------ 4 ------ 1 ------ ad1_address3 --- NULL
5 ------ 5 ------ 1 ------ ad1_ctyfk --- 56
6 ------ 6 ------ 1 ------ ad1_postalcode --- 80155-5036
7 ------ 7 ------ 1 ------ ad1_active --- Y
8 ------ 8 ------ 1 ------ ad1_irstat --- A
9 ------ 9 ------ 1 ------ ad1_irdata --- NULL
10 ------ 10 ------ 1 ------ ad1_at1fk --- 1
1 ------ 11 ------ 2 ------ ad1_pk --- 2
2 ------ 12 ------ 2 ------ ad1_address1 --- 1871 S. Broadway
3 ------ 13 ------ 2 ------ ad1_address2 --- NULL
4 ------ 14 ------ 2 ------ ad1_address3 --- NULL
5 ------ 15 ------ 2 ------ ad1_ctyfk --- 1
6 ------ 16 ------ 2 ------ ad1_postalcode --- 80210
7 ------ 17 ------ 2 ------ ad1_active --- Y
8 ------ 18 ------ 2 ------ ad1_irstat --- A
9 ------ 19 ------ 2 ------ ad1_irdata --- NULL
10 ------ 20 ------ 2 ------ ad1_at1fk --- 1
1 ------ 21 ------ 3 ------ ad1_pk --- 3
2 ------ 22 ------ 3 ------ ad1_address1 --- 1575 W. 124th Avenue
3 ------ 23 ------ 3 ------ ad1_address2 --- NULL
4 ------ 24 ------ 3 ------ ad1_address3 --- NULL
5 ------ 25 ------ 3 ------ ad1_ctyfk --- 1
6 ------ 26 ------ 3 ------ ad1_postalcode --- 80234
7 ------ 27 ------ 3 ------ ad1_active --- Y
8 ------ 28 ------ 3 ------ ad1_irstat --- A
9 ------ 29 ------ 3 ------ ad1_irdata --- NULL
10 ------ 30 ------ 3 ------ ad1_at1fk --- 1
1 ------ 31 ------ 4 ------ ad1_pk --- 4
2 ------ 32 ------ 4 ------ ad1_address1 --- 6175 S. Forest Court
3 ------ 33 ------ 4 ------ ad1_address2 --- NULL
4 ------ 34 ------ 4 ------ ad1_address3 --- NULL
5 ------ 35 ------ 4 ------ ad1_ctyfk --- 101
6 ------ 36 ------ 4 ------ ad1_postalcode --- 80121
7 ------ 37 ------ 4 ------ ad1_active --- Y
8 ------ 38 ------ 4 ------ ad1_irstat --- A
9 ------ 39 ------ 4 ------ ad1_irdata --- NULL
10 ------ 40 ------ 4 ------ ad1_at1fk --- 1
1 ------ 41 ------ 5 ------ ad1_pk --- 5
2 ------ 42 ------ 5 ------ ad1_address1 --- 2000 E. County Line Rd., C-127
3 ------ 43 ------ 5 ------ ad1_address2 --- NULL
4 ------ 44 ------ 5 ------ ad1_address3 --- NULL
5 ------ 45 ------ 5 ------ ad1_ctyfk --- 87
6 ------ 46 ------ 5 ------ ad1_postalcode --- 80126
7 ------ 47 ------ 5 ------ ad1_active --- Y
8 ------ 48 ------ 5 ------ ad1_irstat --- A
9 ------ 49 ------ 5 ------ ad1_irdata --- NULL
10 ------ 50 ------ 5 ------ ad1_at1fk --- 1
1 ------ 51 ------ 6 ------ ad1_pk --- 6
2 ------ 52 ------ 6 ------ ad1_address1 --- 6911 S. University Blvd
3 ------ 53 ------ 6 ------ ad1_address2 --- NULL
4 ------ 54 ------ 6 ------ ad1_address3 --- NULL
5 ------ 55 ------ 6 ------ ad1_ctyfk --- 101
6 ------ 56 ------ 6 ------ ad1_postalcode --- 80122
7 ------ 57 ------ 6 ------ ad1_active --- Y
8 ------ 58 ------ 6 ------ ad1_irstat --- A
9 ------ 59 ------ 6 ------ ad1_irdata --- NULL
10 ------ 60 ------ 6 ------ ad1_at1fk --- 1
1 ------ 61 ------ 7 ------ ad1_pk --- 7
2 ------ 62 ------ 7 ------ ad1_address1 --- 215 Union Blvd., Suite 350
3 ------ 63 ------ 7 ------ ad1_address2 --- NULL
4 ------ 64 ------ 7 ------ ad1_address3 --- NULL
5 ------ 65 ------ 7 ------ ad1_ctyfk --- 96
6 ------ 66 ------ 7 ------ ad1_postalcode --- 80228
7 ------ 67 ------ 7 ------ ad1_active --- Y
8 ------ 68 ------ 7 ------ ad1_irstat --- A
9 ------ 69 ------ 7 ------ ad1_irdata --- NULL
10 ------ 70 ------ 7 ------ ad1_at1fk --- 1
1 ------ 71 ------ 8 ------ ad1_pk --- 8
2 ------ 72 ------ 8 ------ ad1_address1 --- 9700 E. Powers Avenue
3 ------ 73 ------ 8 ------ ad1_address2 --- NULL
4 ------ 74 ------ 8 ------ ad1_address3 --- NULL
5 ------ 75 ------ 8 ------ ad1_ctyfk --- 56
6 ------ 76 ------ 8 ------ ad1_postalcode --- 80111
7 ------ 77 ------ 8 ------ ad1_active --- Y
8 ------ 78 ------ 8 ------ ad1_irstat --- A
9 ------ 79 ------ 8 ------ ad1_irdata --- NULL
10 ------ 80 ------ 8 ------ ad1_at1fk --- 1
1 ------ 81 ------ 9 ------ ad1_pk --- 9
2 ------ 82 ------ 9 ------ ad1_address1 --- 650 S. Cherry Street, #420
3 ------ 83 ------ 9 ------ ad1_address2 --- NULL
4 ------ 84 ------ 9 ------ ad1_address3 --- NULL
5 ------ 85 ------ 9 ------ ad1_ctyfk --- 1
6 ------ 86 ------ 9 ------ ad1_postalcode --- 80222
7 ------ 87 ------ 9 ------ ad1_active --- Y
8 ------ 88 ------ 9 ------ ad1_irstat --- A
9 ------ 89 ------ 9 ------ ad1_irdata --- NULL
10 ------ 90 ------ 9 ------ ad1_at1fk --- 1
1 ------ 91 ------ 10 ------ ad1_pk --- 10
2 ------ 92 ------ 10 ------ ad1_address1 --- 1001 17th St.
3 ------ 93 ------ 10 ------ ad1_address2 --- Suite 1800
4 ------ 94 ------ 10 ------ ad1_address3 --- NULL
5 ------ 95 ------ 10 ------ ad1_ctyfk --- 1
6 ------ 96 ------ 10 ------ ad1_postalcode --- 80202
7 ------ 97 ------ 10 ------ ad1_active --- Y
8 ------ 98 ------ 10 ------ ad1_irstat --- A
9 ------ 99 ------ 10 ------ ad1_irdata --- NULL
10 ------ 100 ------ 10 ------ ad1_at1fk --- 1
而这最终是我想要的：
ad1_pk --- ad1_address1 ------------------------ ad1_address2 ---- ad1_address3 ---- ad1_ctyfk ---- ad1_postalcode ---- ad1_active ---- ad1_irstat ---- ad1_irdata ---- ad1_at1fk
1 --- P.O. Box 5036 --- NULL ---- NULL ---- 56 ---- 80155-5036 ---- Y ---- A ---- NULL ---- 1
2 --- 1871 S. Broadway --- NULL ---- NULL ---- 1 ---- 80210 ---- Y ---- A ---- NULL ---- 1
3 --- 1575 W. 124th Avenue --- NULL ---- NULL ---- 1 ---- 80234 ---- Y ---- A ---- NULL ---- 1
4 --- 6175 S. Forest Court --- NULL ---- NULL ---- 101 ---- 80121 ---- Y ---- A ---- NULL ---- 1
5 --- 2000 E. County Line Rd., C-127 --- NULL ---- NULL ---- 87 ---- 80126 ---- Y ---- A ---- NULL ---- 1
6 --- 6911 S. University Blvd --- NULL ---- NULL ---- 101 ---- 80122 ---- Y ---- A ---- NULL ---- 1
7 --- 215 Union Blvd., Suite 350 --- NULL ---- NULL ---- 96 ---- 80228 ---- Y ---- A ---- NULL ---- 1
8 --- 9700 E. Powers Avenue --- NULL ---- NULL ---- 56 ---- 80111 ---- Y ---- A ---- NULL ---- 1
9 --- 650 S. Cherry Street, #420 --- NULL ---- NULL ---- 1 ---- 80222 ---- Y ---- A ---- NULL ---- 1
10 --- 1001 17th St. --- Suite 1800 ---- NULL ---- 1 ---- 80202 ---- Y ---- A ---- NULL ---- 1
这可能吗？
=============================================== =========================
**编辑：透视表的可能解决方案**
DECLARE @thisPK int
DECLARE @thisDRN INT
DECLARE @thisDRN2 INT
DECLARE @thisDRN3 int
DECLARE @thisIDStart INT
DECLARE @thisIDEnd INT
DECLARE @thisNumIDsAffected int
DECLARE @thisID int
DECLARE @thisColumn VARCHAR(255)
DECLARE @thisValue VARCHAR(255)
--DECLARE @thisRecNum VARCHAR(MAX) = 1 -- the record indicator can be a char(1) or an int
DECLARE @distinctColumnsCount INT = (SELECT COUNT(*) [Cnt] FROM (SELECT DISTINCT Name FROM dbo.zTempXMLDataConverted2) Step1 )
DECLARE @currentImportTable VARCHAR(MAX)='imaddad1'
DECLARE @SQL VARCHAR(MAX)
DECLARE @SQLcolumnNames VARCHAR(MAX)
DECLARE @SQLcolumnValues VARCHAR(MAX)
DECLARE recordGroupCursor CURSOR FOR (
SELECT DENSE_RANK() OVER (ORDER BY NumIDsAffected) drn,NumIDsAffected FROM (SELECT DISTINCT NumIDsAffected FROM #tempUpdateRanges) t1
WHERE NumIDsAffected=9
)
OPEN recordGroupCursor
FETCH NEXT FROM recordGroupCursor INTO @thisDRN,@thisNumIDsAffected
WHILE @@FETCH_STATUS=0
BEGIN
SET @SQL='insert into '+@currentImportTable+' '
SET @SQLcolumnNames='('
SET @SQLcolumnValues=''
DECLARE recordPullCursor CURSOR FOR (
SELECT DENSE_RANK() OVER (ORDER BY PK) drn,PK,IDStart,IDEnd FROM #tempUpdateRanges
WHERE NumIDsAffected=@thisNumIDsAffected
)
OPEN recordPullCursor
FETCH NEXT FROM recordPullCursor INTO @thisDRN2,@thisPK,@thisIDStart,@thisIDEnd--,@thisNumIDsAffected
WHILE @@FETCH_STATUS=0 --AND @thisDRN2<10
BEGIN
--if there happens to be any anomalies in the data, where there are more columns present in a record
--then the distinct column count, then only get the number of columns that can fit into the table
IF @thisNumIDsAffected>@distinctColumnsCount BEGIN
SET @thisIDEnd = @thisIDStart+@distinctColumnsCount
END
SET @SQLcolumnValues=@SQLcolumnValues
DECLARE recordDetailsCursor CURSOR FOR (
SELECT DENSE_RANK() OVER (ORDER BY ID) drn,ID,Name,Value FROM dbo.zTempXMLDataConverted2 WHERE ID between @thisIDStart AND @thisIDEnd
)
OPEN recordDetailsCursor
FETCH NEXT FROM recordDetailsCursor INTO @thisDRN3,@thisID,@thisColumn,@thisValue
WHILE @@FETCH_STATUS=0
BEGIN
IF @thisDRN2=1 BEGIN
SET @SQLcolumnNames += @thisColumn+','+CHAR(10)
END
SET @SQLcolumnValues += ''''+ISNULL(@thisValue,'')+''','
FETCH NEXT FROM recordDetailsCursor INTO @thisDRN3,@thisID,@thisColumn,@thisValue
END
CLOSE recordDetailsCursor
DEALLOCATE recordDetailsCursor
SET @SQLcolumnValues = LEFT(@SQLcolumnValues,LEN(@SQLcolumnValues)-1)+' UNION ALL '+CHAR(10)+'SELECT '
--PRINT @SQLcolumnValues
FETCH NEXT FROM recordPullCursor INTO @thisDRN2,@thisPK,@thisIDStart,@thisIDEnd--,@thisNumIDsAffected
END
CLOSE recordPullCursor
DEALLOCATE recordPullCursor
SET @SQL += LEFT(@SQLcolumnNames,LEN(@SQLcolumnNames)-2)+') '+CHAR(10)+'SELECT '+LEFT(@SQLcolumnValues,LEN(@SQLcolumnValues)-11)
BEGIN TRY
--SET @SQLerrors=''
--PRINT @SQL
EXEC (@SQL)
--RAISERROR(' Record %10i Processed',0,1,@thisPK) WITH NOWAIT
END TRY
BEGIN CATCH
--SET @SQLerrors = 'ERROR!: Creating table '+@currentImportTable+CHAR(10)+ERROR_MESSAGE()
PRINT @SQL+CHAR(10)+ERROR_MESSAGE()
--GOTO EARLY_END
END CATCH
FETCH NEXT FROM recordGroupCursor INTO @thisDRN,@thisNumIDsAffected
END
CLOSE recordGroupCursor
DEALLOCATE recordGroupCursor
SET NOCOUNT OFF;
答案 0 :(得分：1)
为了让drn重新开始时更改RecNum，你可以使用这种逻辑：
select
drn, id,
dense_rank() over (order by GRP) as RecNum
from (
select drn, id, id-drn as GRP from table1
) X
order by id
这将使用drn和id之间的差异来确定哪个＆＃34;组＆＃34;该行属于并使用密集排名来对行进行编号。您可以在SQL Fiddle中查看此内容。假设数据中没有跳跃/间隙或任何奇怪的内容，否则RecNum也会跳跃。
根据编辑，我会假设您的最终专栏未修复，因此您需要进行动态调整。在SO中有很多例子，例如here。