XPath sorting not persistent?

Question

I have a following XML:

<doc>
<ActivityNarrativeInformation>
  <ActivityID>123456789</ActivityID>
  <ActivityNarrativeInformationID>111111111</ActivityNarrativeInformationID>
  <ActivityNarrativeSequenceNumber>1</ActivityNarrativeSequenceNumber>
  <ActivityNarrativeText>She Sells Sea Shells by the Sea Shore and she also</ActivityNarrativeText>
  </ActivityNarrativeInformation>
 <ActivityNarrativeInformation>
  <ActivityID>123456789</ActivityID>
  <ActivityNarrativeInformationID>111111111</ActivityNarrativeInformationID>
  <ActivityNarrativeSequenceNumber>3</ActivityNarrativeSequenceNumber>
  <ActivityNarrativeText>triple shot frappuccino, extra hot, with whipped cream in a tall cup </ActivityNarrativeText>
</ActivityNarrativeInformation>
<ActivityNarrativeInformation>
  <ActivityID>123456789</ActivityID>
  <ActivityNarrativeInformationID>111111111</ActivityNarrativeInformationID>
  <ActivityNarrativeSequenceNumber>2</ActivityNarrativeSequenceNumber>
  <ActivityNarrativeText>likes to take long walks on the beach while she drinks a</ActivityNarrativeText>
  </ActivityNarrativeInformation>
<ActivityNarrativeInformation>
  <ActivityID>987654321</ActivityID>
  <ActivityNarrativeInformationID>222222222</ActivityNarrativeInformationID>
  <ActivityNarrativeSequenceNumber>486</ActivityNarrativeSequenceNumber>
  <ActivityNarrativeText>It was a dark and stormy night; the rain fell in torrents--except at occasional intervals, when
 </ActivityNarrativeText>
</ActivityNarrativeInformation>
<ActivityNarrativeInformation>
  <ActivityID>987654321</ActivityID>
  <ActivityNarrativeInformationID>222222222</ActivityNarrativeInformationID>
  <ActivityNarrativeSequenceNumber>488</ActivityNarrativeSequenceNumber>
  <ActivityNarrativeText>scene lies), rattling along the housetops, and fiercely agitating the scanty flame of the lamps that
</ActivityNarrativeText>
</ActivityNarrativeInformation>
<ActivityNarrativeInformation>
  <ActivityID>987654321</ActivityID>
  <ActivityNarrativeInformationID>222222222</ActivityNarrativeInformationID>
  <ActivityNarrativeSequenceNumber>487</ActivityNarrativeSequenceNumber>
  <ActivityNarrativeText>was checked by a violent gust of wind which swept up the streets (for it is in London that our
</ActivityNarrativeText>
</ActivityNarrativeInformation>
<ActivityNarrativeInformation>
  <ActivityID>987654321</ActivityID>
  <ActivityNarrativeInformationID>222222222</ActivityNarrativeInformationID>
  <ActivityNarrativeSequenceNumber>489</ActivityNarrativeSequenceNumber>
  <ActivityNarrativeText>struggled against the darkness.
</ActivityNarrativeText>
</ActivityNarrativeInformation>
<ActivityNarrativeInformation>
  <ActivityID>55555555</ActivityID>
  <ActivityNarrativeInformationID>77777777</ActivityNarrativeInformationID>
  <ActivityNarrativeSequenceNumber>31921</ActivityNarrativeSequenceNumber>
  <ActivityNarrativeText>Papa Bear was very big and growly. Mamma Bear was middle-sized and pleasant.
</ActivityNarrativeText>
</ActivityNarrativeInformation>
<ActivityNarrativeInformation>
  <ActivityID>55555555</ActivityID>
  <ActivityNarrativeInformationID>77777777</ActivityNarrativeInformationID>
  <ActivityNarrativeSequenceNumber>31923</ActivityNarrativeSequenceNumber>
  <ActivityNarrativeText>Papa bear loved to fix things around the house; Mama bear loved to grow flowers in her garden; and, Baby bear loved playing in the yard. They were very happy. </ActivityNarrativeText>
</ActivityNarrativeInformation>
<ActivityNarrativeInformation>
  <ActivityID>55555555</ActivityID>
  <ActivityNarrativeInformationID>77777777</ActivityNarrativeInformationID>
  <ActivityNarrativeSequenceNumber>31920</ActivityNarrativeSequenceNumber>
  <ActivityNarrativeText>Once upon a time there were three bears, Papa Bear, Mamma Bear and Baby Bear
</ActivityNarrativeText>
</ActivityNarrativeInformation>
<ActivityNarrativeInformation>
  <ActivityID>55555555</ActivityID>
  <ActivityNarrativeInformationID>77777777</ActivityNarrativeInformationID>
  <ActivityNarrativeSequenceNumber>31922</ActivityNarrativeSequenceNumber>
  <ActivityNarrativeText>And Baby Bear, well, he was small, and
sometimes he squeaked! They lived in a pretty little house on the edge of the forest
</ActivityNarrativeText>
</ActivityNarrativeInformation>
</doc

I need to group ActivityNarrativeInformation elements by ActivityID and concatenate ActivityNarrativeText in such a way, that it is sorted by ActivityNarrativeSequenceNumber

I managed to sort elements with following XPath query (XPath 3.1) sort(//ActivityNarrativeInformation[ActivityID=123456789], (), function($ActivityNarrativeSequenceNumber) {$ActivityNarrativeSequenceNumber})

So the result looks like this:

<ActivityNarrativeInformation>
  <ActivityID>123456789</ActivityID>
  <ActivityNarrativeInformationID>111111111</ActivityNarrativeInformationID>
  <ActivityNarrativeSequenceNumber>1</ActivityNarrativeSequenceNumber>
  <ActivityNarrativeText>She Sells Sea Shells by the Sea Shore and she also</ActivityNarrativeText>
  </ActivityNarrativeInformation>
<ActivityNarrativeInformation>
  <ActivityID>123456789</ActivityID>
  <ActivityNarrativeInformationID>111111111</ActivityNarrativeInformationID>
  <ActivityNarrativeSequenceNumber>2</ActivityNarrativeSequenceNumber>
  <ActivityNarrativeText>likes to take long walks on the beach while she drinks a</ActivityNarrativeText>
  </ActivityNarrativeInformation>
<ActivityNarrativeInformation>
  <ActivityID>123456789</ActivityID>
  <ActivityNarrativeInformationID>111111111</ActivityNarrativeInformationID>
  <ActivityNarrativeSequenceNumber>3</ActivityNarrativeSequenceNumber>
  <ActivityNarrativeText>triple shot frappuccino, extra hot, with whipped cream in a tall cup </ActivityNarrativeText>
</ActivityNarrativeInformation>

The probelm however is, that if I want to limit down above to just all ActivityNarrativeText by adding /ActivityNarrativeText at the end like this

sort(//ActivityNarrativeInformation[ActivityID=123456789], (), function($ActivityNarrativeSequenceNumber) {$ActivityNarrativeSequenceNumber})/ActivityNarrativeText

or

sort(//ActivityNarrativeInformation[ActivityID=123456789]/ActivityNarrativeText, (), function($seq) {$seq/ActivityNarrativeSequenceNumber})

The order is lost:

<ActivityNarrativeText>She Sells Sea Shells by the Sea Shore and she also</ActivityNarrativeText>
<ActivityNarrativeText>triple shot frappuccino, extra hot, with whipped cream in a tall cup </ActivityNarrativeText>
<ActivityNarrativeText>likes to take long walks on the beach while she drinks a</ActivityNarrativeText>

What am I doing wrong?

Answer 1

In addition to the right answer not to use / after sorting but ! instead, one of your attempts would actually work if your sort function argument selected the right element as the sort key:

sort(//ActivityNarrativeInformation[ActivityID=123456789]/ActivityNarrativeText, (), function($text) {$text/../ActivityNarrativeSequenceNumber})

Answer 2

You lose the order when you write /ActivityNarrativeText, and it returns the <ActivityNarrativeText> in the same order they have in the input file

/something with nodes does not just mean map it to the child.

It means

• Map it

• Reorder all nodes to the input document order

• Remove duplicates

You could use !ActivityNarrativeText

Answer 3

Testing it here: videlibri.de/cgi-bin/xidelcgi

If you're using xidel, then please add its tag. And maybe cmd for Windows, or bash for Unix as well.

I'm not too sure this can be done with XPath. I believe you're better off using XQuery.

For the narrative with <ActivityID>123456789</ActivityID> you could do:

$ xidel -s input.xml --xquery '
  normalize-space(
    for $x in //ActivityNarrativeInformation
    where $x/ActivityID = 123456789
    order by $x/ActivityNarrativeSequenceNumber
    return
    $x/ActivityNarrativeText
  )
'

For all narratives I'd suggest:

$ xidel -s input.xml --xquery '
  for $narrative at $i in //ActivityNarrativeInformation
  group by $id:=$narrative/ActivityID
  count $i
  return (
    $i,
    normalize-space(
      for $seq in $narrative
      order by $seq/ActivityNarrativeSequenceNumber
      return
      $seq/ActivityNarrativeText
    )
  )
'
1
Once upon a time there were three bears, [...]
2
She Sells Sea Shells by the Sea Shore and [...]
3
It was a dark and stormy night; the rain [...]

Group by <ActivityID> first, then in another for-loop order the sentences by <ActivityNarrativeSequenceNumber>.

Update 2021-07-05; I forgot about XPath's !. In that case one for-loop is enough:

$ xidel -s input.xml --xquery '
  for $narrative at $i in //ActivityNarrativeInformation
  order by $narrative/ActivityNarrativeSequenceNumber
  group by $id:=$narrative/ActivityID
  count $i
  return (
    $i,
    normalize-space($narrative ! ActivityNarrativeText)
  )
'

Answer 4

If what you want to do is extract a coherenet sentece from your sample xml from that particular ActivityID, this expression

string-join(sort(//ActivityNarrativeInformation[ActivityID=123456789]/ActivityNarrativeText/concat(normalize-space()," "), (), function($ActivityNarrativeSequenceNumber) {$ActivityNarrativeSequenceNumber}))

should output

She Sells Sea Shells by the Sea Shore and she also likes to take long walks on the beach while she drinks a triple shot frappuccino, extra hot, with whipped cream in a tall cup