如果找不到多个实例，则强制 CharIndex 仅返回 0？答案

【问题标题】：Force CharIndex return only 0 if more than one instance is not found?如果找不到多个实例，则强制 CharIndex 仅返回 0？
【发布时间】：2020-06-14 12:54:09
【问题描述】：

我的 Row_1 值为

<FeeSize><XMLs><XML Character="1.11" MaxXML="249999.99" Fee="1.60" Cost="1.8" Size="1.8"/><XML Character="251111.11" MaxXML="499999.99" Fee="1.40" Cost="1.7" Size="1.7"/><XML Character="511111.11" MaxXML="999999.99" Fee="1.20" Cost="1.6" Size="1.6"/><XML Character="1111111.11" MaxXML="1999999.99" Fee="1.11" Cost="1.5" Size="1.5"/><XML Character="2111111.11" MaxXML="4999999.99" Fee="1.90" Cost="1.45" Size="1.45"/><XML Character="5111111.11" MaxXML="9999999999999.99" Fee="1.85" Cost="1.425" Size="1.425"/></XMLs></FeeSize>

我的 Row_2 值为

<FeeSize><XMLs><XML Character="1.11" MaxXML="249999.99" Fee="1.60" Cost="1.8" Size="1.8"/></XMLs></FeeSize>

我的要求是识别记录中字符串“Character”的所有起始位置。

我使用的查询是

CHARINDEX('Character', Record, 1) First_C,
CHARINDEX('Character', Record, CHARINDEX('Character', Record)+1) Second_C,
CHARINDEX('Character', Record, CHARINDEX('Character', Record,CHARINDEX('Character', Record)+1)+1) Third_C,
CHARINDEX('Character', Record, CHARINDEX('Character', Record,CHARINDEX('Character', Record, CHARINDEX('Character', Record)+1)+1)+1) Fourth_C,
CHARINDEX('Character', Record, CHARINDEX('Character', Record,CHARINDEX('Character', Record, CHARINDEX('Character', Record,CHARINDEX('Character', Record)+1)+1)+1)+1) Fifth_C,
CHARINDEX('Character', Record, CHARINDEX('Character', Record,CHARINDEX('Character', Record, CHARINDEX('Character', Record,CHARINDEX('Character', Record, CHARINDEX('Character', Record)+1)+1)+1)+1)+1) Sixth_C

现在我可以从 Row_1 看到“字符”的所有起始位置。

由于 Row_2 只有一个“Character”实例，我应该只在第一列 First_C 获得位置。但是，我在第三列 Third_C 和第五列 Fifth_C 获得了第一个起始位置值，即使不应该有任何值。

如何为 Row_2 的第二到六列获取“0”？

【问题讨论】：

标签： sql-server database function sql-server-2005

【解决方案1】：

我确信有更好的方法可以做到这一点，但以下代码将为您提供所需的结果。

Use AdventureWorks2017;
Go




Declare @xml1 nvarchar(max)
Set @xml1 ='<FeeSize><XMLs><XML Character="1.11" MaxXML="249999.99" Fee="1.60" Cost="1.8" Size="1.8"/><XML Character="251111.11" MaxXML="499999.99" Fee="1.40" Cost="1.7" Size="1.7"/><XML Character="511111.11" MaxXML="999999.99" Fee="1.20" Cost="1.6" Size="1.6"/><XML Character="1111111.11" MaxXML="1999999.99" Fee="1.11" Cost="1.5" Size="1.5"/><XML Character="2111111.11" MaxXML="4999999.99" Fee="1.90" Cost="1.45" Size="1.45"/><XML Character="5111111.11" MaxXML="9999999999999.99" Fee="1.85" Cost="1.425" Size="1.425"/></XMLs></FeeSize>'

Declare @xml2 nvarchar(max)
Set @xml2 = '<FeeSize><XMLs><XML Character="1.11" MaxXML="249999.99" Fee="1.60" Cost="1.8" Size="1.8"/></XMLs></FeeSize>'

If Object_ID ('tempdb..#Six') is not null Drop Table #Six
Create Table #Six (id int identity (1,1), xml_1 nvarchar(max), xml_2 nvarchar(max))
Insert into #Six (xml_1, xml_2)
Select @xml1, @xml2 UNION ALL
Select @xml1, @xml2 UNION ALL
Select @xml1, @xml2 UNION ALL
Select @xml1, @xml2 UNION ALL
Select @xml1, @xml2 UNION ALL
Select @xml1, @xml2 


Select case 
when id = 1 then charindex('Character',xml_1,1)
when id = 2 and charindex('Character',xml_1,1) <> 0
then charindex('Character',xml_1,charindex('Character',xml_1,1)+1)
when id = 3 and charindex('Character',xml_1,charindex('Character',xml_1,1)+1) <> 0
then charindex('Character',xml_1,charindex('Character',xml_1,charindex('Character',xml_1,charindex('Character',xml_1,1)+1)+1)+1)
when id = 4 and charindex('Character',xml_1,charindex('Character',xml_1,charindex('Character',xml_1,charindex('Character',xml_1,1)+1)+1)+1) <> 0
then charindex('Character',xml_1,charindex('Character',xml_1,charindex('Character',xml_1,charindex('Character',xml_1,charindex('Character',xml_1,1)+1)+1)+1)+1)
when id = 5 and charindex('Character',xml_1,charindex('Character',xml_1,charindex('Character',xml_1,charindex('Character',xml_1,charindex('Character',xml_1,1)+1)+1)+1)+1) <> 0
then charindex('Character',xml_1,charindex('Character',xml_1,charindex('Character',xml_1,charindex('Character',xml_1,charindex('Character',xml_1,charindex('Character',xml_1,1)+1)+1)+1)+1)+1)
when id = 6 and charindex('Character',xml_1,charindex('Character',xml_1,charindex('Character',xml_1,charindex('Character',xml_1,charindex('Character',xml_1,charindex('Character',xml_1,1)+1)+1)+1)+1)+1) <> 0
then charindex('Character',xml_1,charindex('Character',xml_1,charindex('Character',xml_1,charindex('Character',xml_1,charindex('Character',xml_1,charindex('Character',xml_1,charindex('Character',xml_1,1)+1)+1)+1)+1)+1)+1)
ELSE 0 end as cascaded_xml1,
case 
when id = 1 then charindex('Character',xml_2,1)
when id = 2 and charindex('Character',xml_2,1) <> 0
then charindex('Character',xml_2,charindex('Character',xml_2,1)+1)
when id = 3 and charindex('Character',xml_2,charindex('Character',xml_2,1)+1) <> 0
then charindex('Character',xml_2,charindex('Character',xml_2,charindex('Character',xml_2,charindex('Character',xml_2,1)+1)+1)+1)
when id = 4 and charindex('Character',xml_2,charindex('Character',xml_2,charindex('Character',xml_2,charindex('Character',xml_2,1)+1)+1)+1) <> 0
then charindex('Character',xml_2,charindex('Character',xml_2,charindex('Character',xml_2,charindex('Character',xml_2,charindex('Character',xml_2,1)+1)+1)+1)+1)
when id = 5 and charindex('Character',xml_2,charindex('Character',xml_2,charindex('Character',xml_2,charindex('Character',xml_2,charindex('Character',xml_2,1)+1)+1)+1)+1) <> 0
then charindex('Character',xml_2,charindex('Character',xml_2,charindex('Character',xml_2,charindex('Character',xml_2,charindex('Character',xml_2,charindex('Character',xml_2,1)+1)+1)+1)+1)+1)
when id = 6 and charindex('Character',xml_2,charindex('Character',xml_2,charindex('Character',xml_2,charindex('Character',xml_2,charindex('Character',xml_2,charindex('Character',xml_2,1)+1)+1)+1)+1)+1) <> 0
then charindex('Character',xml_2,charindex('Character',xml_2,charindex('Character',xml_2,charindex('Character',xml_2,charindex('Character',xml_2,charindex('Character',xml_2,charindex('Character',xml_2,1)+1)+1)+1)+1)+1)+1)
ELSE 0 end as cascaded_xml2
from #Six

您的代码中发生的情况是级联的 charindex 被传入，它实际上是 = 0，这使 SQL 回到字符串的开头附近，准备好查找“Character”的新实例。这是相当困难的。上面的代码可以工作，但我确信有更好的方法。

【讨论】：