How to select all nodes such that their group size is higher than a given value, in XPath

Question

I would like to select all <mynode> elements that have a value that appears a certain number of times (say, x) in all the elements.

Example:

<root>
  <mynode>
    <attr1>value_1</attr1>
    <attr2>value_2</attr2>
  </mynode>
  <mynode>
    <attr1>value_3</attr1>
    <attr2>value_3</attr2>
  </mynode>
  <mynode>
    <attr1>value_4</attr1>
    <attr2>value_5</attr2>
  </mynode>
  <mynode>
    <attr1>value_6</attr1>
    <attr2>value_5</attr2>
  </mynode>
</root>

In this case, I want all the <mynode> elements that whose attr2 value occurs > 1 time (x = 1). So, the last two <mynode>s.

Which query I have to perform in order to achieve this target?

I don't get it. Why _the last two nodes_. So I guess you refer to `mynode` at position 3 and 4. However, for both of them the attr_2 value (value_5) just appears one time in other nodes. So the example you propose does not fit to your question (or maybe I am missing or misunderstanding something) — dirkk, Jun 17 '13 at 16:17
Yes, I meant nodes (mynode) 3 and 4. The value_5 of attr2 appears 2 times in *all* the nodes. — Baduel, Jun 17 '13 at 16:21
@Baduel, your question says "that appear a number of times (says x) in the **other** nodes", not "in **all** nodes". Hence the confusion. — LarsH, Jun 17 '13 at 20:27

score 1 · Answer 1 · edited May 23 '17 at 12:13

1

If you're using XPath 2.0 or greater, then the following will work:

for $value in distinct-values(/root/mynode/attr2)
return
    if (count(/root/mynode[attr2 = $value]) > 1) then
        /root/mynode[attr2 = $value]
    else ()

For a more detailed discussion see: XPath/XSLT nested predicates: how to get the context of outer predicate?

edited May 23 '17 at 12:13

Community

1
1

answered Jun 17 '13 at 20:05

Will Goring

1,040
1
7
18

score 0 · Answer 2 · answered Jun 18 '13 at 22:26

This is also possible in plain XPath 1.0 (also works in newer versions of XPath); and probably easier to read. Think of your problem as you're looking for all <mynode/>s which have an <att2/> node that also occurs before or after the <mynode/>:

//mynode[attr2 = preceding::attr2 or attr2 = following::attr2]

If <att2/> nodes can also accour inside other elements and you do not want to test for those:

//mynode[attr2 = preceding::mynode/attr2 or attr2 = following::mynode/attr2]

How to select all nodes such that their group size is higher than a given value, in XPath

2 Answers2