Question

Here is the XML:

<?xml version="1.0" encoding="UTF-16"?>
<Response>
  <Offers>
    <Car>
      <Group>ECMR</Group>
    </Car>
    <Insurances>
      <Optional>
        <Code>BE</Code>
      </Optional>
      <Optional>
        <Code>BF</Code>
      </Optional>
      <Optional>
        <Code>I</Code>
      </Optional>
      <Optional>
        <Code>LD</Code>
      </Optional>
    </Insurances>
  </Offers>
  <Offers>
    <Car>
      <Group>CDMR</Group>
    </Car>
    <Insurances>
      <Optional>
        <Code>BA</Code>
      </Optional>
      <Optional>
        <Code>BG</Code>
      </Optional>
      <Optional>
        <Code>IS</Code>
      </Optional>
      <Optional>
        <Code>LS</Code>
      </Optional>
    </Insurances>
  </Offers>
</Response>

Using this XML and within a single SELECT statement in a SQL Server stored procedure, I would like to create a result set that looks like this:

Group           Codes
--------------------------
ECMR            BE,BF,I,LD
CDMR            BG,BA,IS,LS

Each record in the result set contains a column for the <Car><Group>group</Group></Car> and another column for the concatenation of each <Insurances><Optional><Code>code</Code></Optional></Insurances>.

Is this possible?

Answer 1

select T.N.value('(Car/Group/text())[1]', 'varchar(10)') as [Group],
       (
         select ','+T2.N.value('(./text())[1]', 'varchar(10)')
         from T.N.nodes('Insurances/Optional/Code') as T2(N)
         for xml path(''),  type
       ).value('substring(./text()[1], 2)', 'varchar(100)') as Codes
from @XML.nodes('/Response/Offers') as T(N)