T-SQL Extract部分xml或nvarchar（max）列匹配模式

Question

I'd like to do something that I think is fairly trivial using T-SQL//SQL Server 2008 R2, but I can't seem to figure out a way.

If I were in Java, C#, C++, whatever, I would do:

1. Find position of first occurrance of '123' in string
2. Execute substring operation from that position getting next 50 characters

So, in SQL Server, I'd basically like:

1. Find all rows where column (X) contains said string (basically a LIKE clause)
2. Return 50 characters from that column starting at the said string's location.

Can I do this somehow? I can cast an XML column to nvarchar(max), do a like operation, and do a substring operation, I don't know how to get the position of the said string in the column in the first place though.

---

Sample content requested in comment

CREATE TABLE SampleTable(xmlData xml);

Pretend the value is in one if SampleTable's xmlData column is as follows. I would like to, for debugging purposes, extract the string from the funny unicode Þ character forward 50 characters (or to the end of the file if that's less than 50).

<RootNode>
    <Row>
        <NestedNode1>
            some text.
        </NestedNode1>

        <NestedNode2>
            123456
        </NestedNode2>

        <NestedNode3>
            Þ Some crazy name with unicode letters. Þ
        </NestedNode3>
    </Row>
</RootNode>

---

Answer 1

Are you looking for CHARINDEX?

;WITH CTE AS(
    SELECT CAST (xmlData as nvarchar(max)) as X
    FROM SampleTable
)
SELECT SUBSTRING(X,CHARINDEX(N'Þ',X),50) as [String]
FROM CTE
WHERE CHARINDEX(N'Þ',X)>0