Question

I work with XQuery to do statistics. I have one document like this :

<tr>
  <td>Element 1</td>
  <td>100</td>
</tr>
<tr>
  <td>Element 2</td>
  <td>80</td>
</tr>
<tr>
  <td>Element 3</td>
  <td>40</td>
</tr>
<tr>
  <td>Element 4</td>
  <td>12</td>
</tr>
<tr>
  <td>Element 5</td>
  <td>8</td>
</tr>

And want to retrieve only a part of this document : I want to have only 80% of the sum of the number of element (The Pareto distribution in fact).

In this case, I have a total of 240 elements. I want in my output the first elements so that the total of the elements is 192 (240*80/100).

In this example, the ideal output would have only the first three elements, like this :

<tr>
  <td>Element 1</td>
  <td>100</td>
</tr>
<tr>
  <td>Element 2</td>
  <td>80</td>
</tr>
<tr>
  <td>Element 3</td>
  <td>40</td>
</tr>

I hope I’m clear :s. I am looking for long time, without success, I don’t find how to do…

Thank so much

Answer 1

Use:

   for $total in sum(/*/*/td[2]),
        $pareto in $total*80 div 100,
        $i in 1 to count(/*/*)
      return
        if(sum(/*/*[position() le $i]/td[2]) ge $pareto
         and
           sum(/*/*[position() lt $i]/td[2]) lt $pareto
           )
           then /*/*[position() le $i]
           else ()

When this XPath expression (yes this is an XQuery expression that is also an XPath 2.0 expression) is evaluated against the provided XML (wrapped into a single tope element to be made a well-formed XML document):

<table>
    <tr>
        <td>Element 1</td>
        <td>100</td>
    </tr>
    <tr>
        <td>Element 2</td>
        <td>80</td>
    </tr>
    <tr>
        <td>Element 3</td>
        <td>40</td>
    </tr>
    <tr>
        <td>Element 4</td>
        <td>12</td>
    </tr>
    <tr>
        <td>Element 5</td>
        <td>8</td>
    </tr>
</table>

the wanted, correct result is produced:

<tr>
    <td>Element 1</td>
    <td>100</td>
  </tr>
<tr>
    <td>Element 2</td>
    <td>80</td>
  </tr>
<tr>
    <td>Element 3</td>
    <td>40</td>
</tr>