使用SQL Server 2016
问题：我从网络订单系统收到一个平面文件。我需要将此数据导入我们的内部邮件订购系统。网络订单文件每个项目的每个订单包含1行。因此，如果订购了12个商品，那么将有12个具有相同订购ID的行。我需要对此进行转换，以便订单项位于同一行。需要注意的是，每个订单每行最多只能有5个订单项。因此，订购的12件物品需要变成3行。在第1行和第2行的末尾，需要有一个继续标记，说明顺序在下一行继续。
网络订单文件示例：
SAMPLE# Row orderid ocustomerid odate ofirstname customer_pmntprofileid itemid numitems unitprice itemamount
1 1 171476 5230 11/14/2018 Scott 1324426745 R22046 1 269.95 269.95
1 2 171476 5230 11/14/2018 Scott 1324426745 R81100 12 7.5 90
3 1 171690 6119 11/14/2018 Joe 1382758196 AF49972 6 39.95 239.7
3 2 171690 6119 11/14/2018 Joe 1382758196 AF49973 4 39.95 159.8
3 3 171690 6119 11/14/2018 Joe 1382758196 AF47970 1 43.95 43.95
3 4 171690 6119 11/14/2018 Joe 1382758196 AF44090 1 43.95 43.95
3 5 171690 6119 11/14/2018 Joe 1382758196 AF44091 1 43.95 43.95
3 6 171690 6119 11/14/2018 Joe 1382758196 AF44092 1 43.95 43.95
3 7 171690 6119 11/14/2018 Joe 1382758196 AF44093 1 43.95 43.95
3 8 171690 6119 11/14/2018 Joe 1382758196 AF44094 1 43.95 43.95
3 9 171690 6119 11/14/2018 Joe 1382758196 AF44114 1 50.95 50.95
3 10 171690 6119 11/14/2018 Joe 1382758196 AF44139 1 46.95 46.95
3 11 171690 6119 11/14/2018 Joe 1382758196 AF44138 1 46.95 46.95
3 12 171690 6119 11/14/2018 Joe 1382758196 AF44066 1 43.95 43.95
4 1 171702 6120 11/14/2018 Tim 1382791806 L81313 5 17.95 89.75
4 2 171702 6120 11/14/2018 Tim 1382791806 L81314 1 30.95 30.95
4 3 171702 6120 11/14/2018 Tim 1382791806 L82039 6 8.95 53.7
4 4 171702 6120 11/14/2018 Tim 1382791806 L82013 1 69.95 69.95
我需要将样本1变成1行，将样本3变成3行，并将样本4变成1行。 需要的输出：
SAMPLE # orderid ocustomerid odate ofirstname customer_pmntprofileid ITEM_ID1 QTY1 PRICE1 AMOUNT1 ITEM_ID2 QTY2 PRICE2 AMOUNT2 ITEM_ID3 QTY3 PRICE3 AMOUNT3 ITEM_ID4 QTY4 PRICE4 AMOUNT4 ITEM_ID5 QTY5 PRICE5 AMOUNT5 Continue
1 171476 5230 11/14/2018 Scott 1324426745 R22046 1 269.95 269.95 R81100 12 7.5 90 N
3 171690 6119 11/14/2018 Joe 1382758196 AF49972 6 39.95 239.7 AF49973 4 39.95 159.8 AF47970 1 43.95 43.95 AF44090 1 43.95 43.95 AF44091 1 43.95 43.95 Y
3 171690 6119 11/14/2018 Joe 1382758196 AF44092 1 43.95 43.95 AF44093 1 43.95 43.95 AF44094 1 43.95 43.95 AF44114 1 50.95 50.95 AF44139 1 46.95 46.95 Y
3 171690 6119 11/14/2018 Joe 1382758196 AF44138 1 46.95 46.95 AF44066 1 43.95 43.95 N
4 171702 6120 11/14/2018 Tim 1382791806 L81313 5 17.95 89.75 L81314 1 30.95 30.95 L82039 6 8.95 53.7 L82013 1 69.95 69.95 N
5 171706 70 11/14/2018 Bruce 1382821080 L84264 1 584.99 584.99 L84409 1 584.99 584.99 L84410 1 584.99 584.99 L84633 1 539.99 539.99 L84634 1 539.99 539.99 Y
答案 0 :(得分：0)
您可以通过条件聚合来做到这一点：
select SAMPLE, orderid, ocustomerid, odate, ofirstname, customer_pmntprofileid,
max(case when seqnum % 5 = 1 then itemid end) as itemid_1,
max(case when seqnum % 5 = 1 then numitems end) as numitems_1,
max(case when seqnum % 5 = 1 then unitprice end) as unitprice_1,
max(case when seqnum % 5 = 1 then itemamount end) as itemamount_1,
max(case when seqnum % 5 = 2 then itemid end) as itemid_2,
max(case when seqnum % 5 = 2 then numitems end) as numitems_2,
max(case when seqnum % 5 = 2 then unitprice end) as unitprice_2,
max(case when seqnum % 5 = 2 then itemamount end) as itemamount_2,
max(case when seqnum % 5 = 3 then itemid end) as itemid_1,
max(case when seqnum % 5 = 3 then numitems end) as numitems_1,
max(case when seqnum % 5 = 3 then unitprice end) as unitprice_1,
max(case when seqnum % 5 = 3 then itemamount end) as itemamount_1,
max(case when seqnum % 5 = 4 then itemid end) as itemid_4,
max(case when seqnum % 5 = 4 then numitems end) as numitems_4,
max(case when seqnum % 5 = 4 then unitprice end) as unitprice_4,
max(case when seqnum % 5 = 4 then itemamount end) as itemamount_4,
max(case when seqnum % 5 = 0 then itemid end) as itemid_5,
max(case when seqnum % 5 = 0 then numitems end) as numitems_5,
max(case when seqnum % 5 = 0 then unitprice end) as unitprice_5,
max(case when seqnum % 5 = 0 then itemamount end) as itemamount_5,
(case when max(seqnum) < max(cnt) then 'Y' else 'N' end) as continuation_flag
from (select wo.*,
row_number() over (partition by orderid order by row) as seqnum,
row_number() over (partition by orderid) as cnt
from web_orders wo
) wo
group by SAMPLE, orderid, ocustomerid, odate, ofirstname, customer_pmntprofileid,
(seqnum - 1) / 5;
答案 1 :(得分：0)
您可以这样做：
WITH
t1 AS (select * FROM mytable WHERE row % 5 = 1),
t2 AS (select * FROM mytable WHERE row % 5 = 2),
t3 AS (select * FROM mytable WHERE row % 5 = 3),
t4 AS (select * FROM mytable WHERE row % 5 = 4),
t5 AS (select * FROM mytable WHERE row % 5 = 0)
SELECT t1.SAMPLE,
ceiling(t1.Row/5.0) AS Row,
t1.orderid,
t1.ocustomerid,
t1.odate,
t1.ofirstname,
t1.customer_pmntprofileid,
t1.itemid AS itemid1,
t1.numitems AS numitems1,
t1.unitprice AS unitprice1,
t1.itemamount AS itemAmount1,
t2.itemid AS itemId2,
t2.numitems AS numitems2,
t2.unitprice AS unitprice2,
t2.itemamount AS itemamount2,
t3.itemid AS itemId3,
t3.numitems AS numitems3,
t3.unitprice AS unitprice3,
t3.itemamount AS itemamount3,
t4.itemid AS itemId4,
t4.numitems AS numitems4,
t4.unitprice AS unitprice4,
t4.itemamount AS itemamount4,
t5.itemid AS itemId5,
t5.numitems AS numitems5,
t5.unitprice AS unitprice5,
t5.itemamount AS itemamount5
FROM t1
LEFT JOIN t2 ON t1.orderid = t2.orderid AND CEILING(t1.row/5.0)=ceiling(t2.row/5.0)
LEFT JOIN t3 ON t1.orderid = t3.orderid AND CEILING(t1.row/5.0)=ceiling(t3.row/5.0)
LEFT JOIN t4 ON t1.orderid = t4.orderid AND CEILING(t1.row/5.0)=ceiling(t4.row/5.0)
LEFT JOIN t5 ON t1.orderid = t5.orderid AND CEILING(t1.row/5.0)=ceiling(t5.row/5.0);