How to select nodes by condition EXCEPT the one with specified index-CodePudding

Based on this question let's say we have XML

<root>
  <a default="False">1</a>
  <a default="True">2</a>
  <a default="False">3</a>
  <a default="False">4</a>
  <a default="True">5</a>
  <a default="False">6</a>
  <a default="False">7</a>
</root>

I can select the first node with @default="True" with

//a[@default="True"][1]  # node with value "2"

the second one with

//a[@default="True"][2]  # node with value "5"

...but if I want to select all except the first one

//a[not(self::a[@default="True"][1])]

1 2 3 4 5 6 7 &extract=//a[not(self::a[@default="True"][1])]&extract-kind=xpath3.1&printed-node-format=html&output-format=adhoc&input-format=auto&dot-notation=unambiguous" rel="nofollow noreferrer">it excludes both node with value "2" and "5"

So what's wrong with the last XPath expression? Why it ignores the index?

CodePudding user response：

To select all except the first, use

//a[@default="True"][position() != 1]

What's wrong with your expression? Well self::a only selects one node, so self::a[1] means exactly the same as self::a.

Incidentally, using //a[1] is usually wrong - it should usually be (//a)[1]. It only works in your case because all the a elements are siblings at the same level.

CodePudding user response：

Try this XPath-1.0 expression:

//a[@default='True'][count(preceding-sibling::a[@default='True'])>0]

This expression selects all a elements with at least one predecessor matching the given criteria.

The explanation why your expression //a[not(self::a[@default="True"][1])] doesn't work is simple: nearly all a elements do not satisfy the condition given. The expression selects all as that are not an a with the attribute @default with the value "True".