XML Schema of Data
This is the XML Schema of our data taken from XBench (see the References section) (TC/SD). It has the structure information. Figure 8 tells us that "dictionary" is the root element. "dictionary" has more than one "e" children. The "hwg" element is the child of "e," and so on.

Querying eXist Using XPath
You can find all of the queries that were used in this article below. The following sections show similar queries taken from XBench. Using and understanding XPath is easier than XQuery, so we'll examine these queries first (see Figure 9).

The right-most icon, which looks like a pair of binoculars, is labeled "Query the database with XPath." This is slightly mislabeled, because we can use the same window for XQuery. In the middle of the "Query Dialog," make sure that you select the context /db/XBench. This is the collection in which we stored our data.

Query1
Return the entry matching the headword "minute." This query can be expressed in XPath as:

XPath:
/dictionary/e[hwg/hw="minute"]
Query1 results
<e id="E2">
    <hwg>
        <hw>minute</hw>
        <pr>1/k96SMVL^2</pr>
        <pos>n.</pos>
    </hwg>
    <et>
        <cr>E143</cr>
        <cr>E180</cr>
        <cr>E530</cr>
        <cr>E308</cr>
        <cr>E215</cr>
        <cr>E298</cr>
        <cr>E294</cr>
    </et>
...

Query2
Find the headword with matching the quotation year "1900." This query can be expressed in XPath as:

XPath:
/dictionary/e[ss/s/qp/q/qd="1900"]/hwg/hw

There are 28 matches for this query. Compilation time is 28 msec, execution time is 890 msec. Total time is less than a second. Compilation time reflects the time spent on reading indexes and applying the appropriate algorithms (more on this topic soon). Execution time is the time needed to access the data (stored in DOM format) and return the parts of it that satisfy the query.

Query2 results
<hw>husbandry</hw>
<hw>supper</hw>
<hw>strand</hw>
<hw>nominated</hw>
<hw>saying</hw>
<hw>coram</hw>
<hw>outwards</hw>
<hw>benches</hw>
<hw>faustuses</hw>
<hw>rhapsody</hw>
<hw>rotten</hw>
<hw>punish</hw>
<hw>favours</hw>
<hw>earth</hw>
<hw>italian</hw>
<hw>waits</hw>
<hw>mention</hw>
<hw>sea</hw>
<hw>compelled</hw>
<hw>rumination</hw>
<hw>outrage</hw>
<hw>liege</hw>
<hw>lifted</hw>
<hw>embrace</hw>
<hw>break</hw>
<hw>profession</hw>
<hw>erecting</hw>
<hw>cinna</hw>

Querying eXist Using XQuery
As I mentioned earlier eXist supports XQuery too. At this section we are going to work on two XQuery examples.

Query3
Return quotation text, separated by one unspecified level, matching headword "opinion." This query can be expressed in XQuery as:

XQuery:
for $ent in /dictionary/e
where $ent/*/hw = "opinion"
return
$ent/ss/s/qp/*/qt

Query3 results
<qt>bravely ruthless courts shall lose daringly since the dino</qt>
<qt>doggedly close dinos about the dinos use always about the foxes.ruthlessly
stealt</qt>
<qt>doggedly brave multipliers can nag quietly on the busy realms?dolphins at
the bravely fluf
<cr>E250</cr>
ironic warhorses besides the enticing,
</qt>
...

Items found numbered 20. Compilation time is 16 msec. Execution time is 2469 msec. Total time spent is less than 2.5 seconds. This is a reasonable time for answering a-not-so-easy query. Notice that this query has two stars (*) in it. Star (*) stands for any element.