AND and ORs with XPath. Querying only within the children of a node, checking all the nodes for children matching the query

Question

Lets say I have an xml structure like this:

<s>
  <a>
    <b id="myid">value1</b>
    <c id="test">val1</c>
    <c id="test2">valX</c>
  </a>
  <g>
    <d id="myid">value2</d>
    <e id="test">val2</e>
  </g>
  <i id="myid">value3</i>
  <j id="test">val3</j>
</s>

Notice the tag names are not important. Only the attribute "id" and the textvalue of the node.

The xpath query will be used to check if the values exists, so I can return the whole XML. This xpath query is done on a whole set of different XML strings returning those who match.

Now I want to use xpath to verify if either:

("myid"="value1" AND "test"="val1") OR ("myid"="value2" AND "test"="val2")

How would I write this as an XPath string? "val1" has to be on the same level in the XML as "value1". The same goes for "val2" with "value2", and "val3" and "value3"

I need to write this in a way that makes it easy to expand upon as this XPath query will be generated by code.

The input query for the generator will be on a AND/OR with "(" and ")". So I need to make a parser to create the XPath from this.

Examples:

q1: ("myid"="value1" AND "test"="val1") OR ("myid"="value2" AND "test"="val2") OR 
 ("myid"="value3" AND "test"="val3")
q2: ("myid"="value1" AND "test"="val1") AND ("myid"="value2" AND "test"="val2") 
q3: "myid"="value1" AND ("test"="val1" OR "test"="valX")

What I need is a way to make sure all the @id checks are performed on the same XML subset. So in this XML the check should either get a match from all the @id's directly below the S-node, or a match from all the @id's directly below the A-node, or a match from all @id's directly below the G-node. (and all other similar nodes)

If any match is done, I can return the XML. I do not know how many levels there are as the xml is very generic with the one rule that each value node has an @id attribute with the value name.

"myid"="value1" AND "test"="val1" should return a match.
"myid"="value2" AND "test"="val2" should return a match.
"myid"="value3" AND "test"="val3" should return a match.
"myid"="value3" AND "test"="val1" should NOT return a match.
"myid"="value2" AND "test"="val3" should NOT return a match.

etc.

This might be simple. But I have not found a way to get this done. Can anyone help me with this. Any hints are appreciated.

I hope I have made this understandable.

How would you write the same where upper or lower case is ignored. Like this?

//*[ *[@id [lower-case(.)="myid"] [lower-case(..)="value1"] ] and *[@id [lower-case(.)="test"] [lower-case(..)="val1"] ] ]

or is there a better way. The values are pre .ToLower()'ed.

Vincent Biragnet Vincent Biragnet · Accepted Answer · 2011-12-05T16:32:28

I think you are looking for an expression like that :

//*[ *[@id ='myid'] = 'value1' and *[@id ='test'] = 'val1']

The result will be all the elements that have at least two children with an id attribute with the following constraints :

one of the children must have an id attribute value that is equal to myid, the value of the element has to be value1.
one of the children must have an id attribute value that is equal to test, the value of the element has to be val1.

Then you can add and or or as you wish. For example :

//*[ (*[@id ='myid'] = 'value1' and *[@id ='test'] = 'val1') or (*[@id ='myid'] = 'value2' and *[@id ='test'] = 'val2')]

AND and ORs with XPath. Querying only within the children of a node, checking all the nodes for children matching the query

1 Answers