Using eXist 101 - eXist: A NoSQL Database and Application Development Platform - page 52

• To show totals and compute percentages for each speaker, we need the full set of
spoken lines in the play:
let $ all-lines := $ play-document // SPEECH / LINE
• We need the lines spoken by each speaker. Assuming the name of the speaker is
in $speaker , we can get these with:
$ speaker-lines := $ play-document // SPEECH [ SPEAKER eq $ speaker ]/ LINE
• Given a sequence of LINE elements in a variable $elms (the sequence of LINE ele‐
ments we retrieved in one of the previous two bullets), how many words are spo‐
ken? A rough but usable approximation for this can be calculated by tokenizing
everything said using whitespace boundaries and couningt/aggregating the
results:
sum ( $ elms ! count ( tokenize (., '\s+' )))
This might need a little further explanation:
— The exclamation mark after the $elms expression is an XQuery 3.0 “bang” or
“simple map” operator (see “The simple map operator” on page 114 ). It per‐
forms the operation on the right for all members of the sequence on the left.
You could have done this in several other (and probably more customary)
ways (e.g., using a FLWOR expression), but this seemed a useful way to intro‐
duce one of the new XQuery 3.0 capabilities.
— The tokenize function tokenizes (breaks up) strings on boundaries given by a
regular expression. The regular expression '\s+' signifies a sequence of
whitespace characters, so that gives us the words (more or less, sometimes
punctuation gets in the way, but let's forget about that for now).
— We're not interested in the words themselves but only in how many there are.
So we simply count them.
— The outer sum function sums the numeric results of what's inside, returning
the total of all words spoken in the elements in $elms .
As a last step, let's put this functionality into a local function (because we're going to
use it twice in our code: once for the full play and once for every speaker):
declare function local:word-count ( $ elms as element ()*) as xs:integer
{
sum ( $ elms ! count ( tokenize (., '\s+' )))
};
Now let's put this all together and create a page that shows us the results of our analy‐
sis (just for Hamlet ) in a table. The code for this is shown in Example 3-7 .