Exclude a grandchild from XML using XPath - sql-server

Question

Trying to exclude a specific grandchild when reading an XML structure using SQL Server 2017.
I have an example structure as follows:
<root>
<a id="parent">
<b id = "child1">
<c id = "grandchild1"/>
<c id = "grandchild2"/>
</b>
<b id = "child2">
<c id = "grandchild3"/>
<c id = "grandchild4"/>
</b>
</a>
</root>
I can exclude a specific child using c[not(contains(#id,"grandchild3"))] when I navigate to for example the node <b id = "child2">.
But how would I exclude that specific grandchild when processing the rest of the XML from the root node?
What i would like to receive, so I can write it to a file, is a structure like this:
<root>
<a id="parent">
<b id = "child1">
<c id = "grandchild1"/>
<c id = "grandchild2"/>
</b>
<b id = "child2">
<c id = "grandchild4"/>
</b>
</a>
</root>
As can be seen grandchild3 is now missing from this structure.

Answer 1

Doing this in a SELECT is surprisingly involved, but fortunately you don't need that. You can make a copy of your xml, modify it with .modify, then use that:
declare #original xml = '<root>
<a id="parent">
<b id = "child1">
<c id = "grandchild1"/>
<c id = "grandchild2"/>
</b>
<b id = "child2">
<c id = "grandchild3"/>
<c id = "grandchild4"/>
</b>
</a>
</root>'
declare #new xml = #original
set #new.modify('delete //c[#id="grandchild3"]')
select #new
gives
<root>
<a id="parent">
<b id="child1">
<c id="grandchild1" />
<c id="grandchild2" />
</b>
<b id="child2">
<c id="grandchild4" />
</b>
</a>
</root>
as desired.
The xpath //c[#id="grandchild3"] means "Any c node (anywhere in the document) with an id attribute with value grandchild3".