Question

我有一个如下所示的查询，我想对end_scan-packing_date列进行计算，我该如何实现？

 SELECT a.invoice_no,
       CONVERT(VARCHAR, b.packing_date,3) as packing_date ,
       CONVERT(VARCHAR, b.exw_date,3) as exw,
       CONVERT(VARCHAR, b.bcd_end_date, 3) as end_date,
       (SELECT TOP 1 insert_date FROM   wms.bcd_shipment d,wms.shinvoicedetails WHERE d.invoice_no = a.invoice_no ORDER  BY insert_date ASC) as start_date ,
       (SELECT TOP 1 insert_date FROM   wms.bcd_shipment d,wms.shinvoicedetails WHERE d.invoice_no = a.invoice_no ORDER  BY insert_date DESC) as end_scan,
       CONVERT(VARCHAR, send_date, 3) as swnd_date,customer_short_name,
       Sum(picking_qty) as qty,total_carton
FROM   wms.shinvoicedetails a,
       wms.shinvoiceheder b,
       wms.shinvoice_ctrl c
WHERE  send_date BETWEEN '2014/8/26 00:00:01' AND '2014/9/25 23:59:59'
       AND a.invoice_no = b.invoice_no
       AND a.invoice_no = c.invoice_no
GROUP  BY a.invoice_no,
          b.packing_date,
          b.exw_date,
          b.bcd_end_date,
          send_date,
          customer_short_name,
          total_carton
ORDER  BY 1;

Answer 1

我可以向您推荐DATEDIFF函数的文档（我假设您将SQL Server用作RDBMS）。

它有3个参数，我假设你想要天数的差异。所以你想要的东西会是这样的：

DATEDIFF(DAY,
         (SELECT TOP 1 insert_date
          FROM   wms.bcd_shipment d,
                 wms.shinvoicedetails
          WHERE  d.invoice_no = a.invoice_no
          ORDER  BY insert_date DESC),
          b.packing_date
) AS diff_days

但您的查询在许多级别上都是错误的。您正在对几乎所有列进行分组以获取SUM，而您应将其编写为a derived table（子查询）仅在必要字段上进行分组。您也转换为VARCHAR，可能转换为VARCHAR（1）。您应该将其写为CONVERT（VARCHAR（10），＆lt; your date field＆gt;，3）。

Answer 2

一天不在00:00:01开始，并且没有以23:59:59结束。帮助自己不要试图将您的需求扩展到between的语法。有一种更简单，更可靠的方法

WHERE  send_date >= '20140826' AND send_date < '20140926'

少于26日;它更准确＆amp;没有愚蠢的一秒扣除。

日期和时间范围的最佳做法是避免BETWEEN和   总是使用表格：

WHERE col＆gt; =＆＃39; 20120101＆＃39; AND col＆lt; ＆＃39; 20120201＆＃39;

此表单适用于所有人   类型和所有精度，无论时间部分是什么   适用。

Itzik Ben-Gan

http://sqlmag.com/t-sql/t-sql-best-practices-part-2

日期减去日期查询

2 个答案: