Example XPaths

When using Saxon you have extensive possibilities using XPath in BlackLab configuration. Some noteworthy examples are shown below.

To learn more about modern XPath, Altova's XPath 3 training is a good resource, and there are many others.

Capture punctuation between words

To capture text content between <w/> tags:

punctPath: .//text()[not(ancestor::w)]

This captures any text node that is not a descendant of a <w/> tag.

Another possible approach:

punctPath: .//text()[.!='' and preceding-sibling::tei:w]|.//tei:pc |.//tei:lb

This captures non-empty text nodes after a <w/> tag plus (the text contents of) pc or lb tags.

Isolate a part of speech feature

Your data may have part of speech information that includes detailed features. Let's say this information is stored in an attribute with values like UPosTag=PRON|Case=Nom|Person=3|PronType=Prs. You can isolate the value of the Case feature like this:

valuePath: replace(./@msd, '.*Case=([A-Za-z0-9]+).*', '$1')

Use default if value is missing

If some of your words have a lemma attribute, and you want to index the value _UNKNOWN_ if it's missing (perhaps to be able to locate these data problems easily), you can do that as follows:

valuePath: ./(string(@lemma), '_UNKNOWN_')[1]

Using either an attribute, or a standoff annotation

Again, let's say some of your words have lemma attributes. But some have the lemma in a separate tei:join element instead. You might use XPath to look up the appropriate value like this:

valuePath: >-
  let $xid := @xml:id
  return if (@lemma) then @lemma 
  else if ($xid) then
    following-sibling::tei:join[@lemma][matches(@target,'#'||$xid||'( |$)')]/@lemma 
  else ()

Note how we can easily split XPath expressions over multiple lines using >- in YAML. (see YAML multiline strings).

For loops

You can even use for loops if necessary, e.g.:

for $w in //tei:w[@xml:id]
return let $xid := $w/@xml:id
    return 
    if ($w/@lemma) then
        $w/@lemma else
            if ($xid) then
                let $join := $w/following-sibling::tei:join[@lemma][matches(@target,concat('#',$xid,'( |$)'))]
                return
                $join/@lemma else
                ()

Thanks to @eduarddrenth for the initial Saxon version and some of the examples.