Navigating the XML Tree

XML representations Blah blah blah ... Wines

White wines ...

Red wines ...

Beers

Ales ...

Lagers ...

stuff ...

Classic boxes-inside-boxes representation of a mythical book that contains an introduction, two chapters, and an index, where each chapter contains a heading and two sections

Classic tree representation of a mythical book that contains an introduction, two chapters, and an index, where each chapter contains a heading and two sections

Sample text Warp Speed, Ms Bright!

Sample Document Instance Warp Speed, Ms Bright! There was a young lady named Bright, Who travelled much faster than light, She departed one day, In a relative way way, And returned on the previous night. See .

Sample Portion of TEI Tree

Here is a graphic representation of the grossly simplified tree representation of the previos TEI sub-document. (It is not a full TEI document, as it does not have a teiHeader.)

Classic simplified tree diagram of the limerick Warp Speed, Ms Bright! that uses a dashed curved arrow to represent the link from the note to the term, and a dotted arror to indicate the link from the ptr to the Wikipedia article.

Walking the Tree

tree (with self::)

Classic tree diagram of generic XML document, elements only, with /*[1]/*[2]/*[2]/*[3] highlighted.

descendants (child:: and descendant::)

Classic tree diagram of generic XML document, elements only, with /*[1]/*[2]/*[2]/*[3] highlighted, all its children marked in green, and all its descendants (including the children) surrounded with a green circle.

ancestors (parent:: and ancestor::)

Classic tree diagram of generic XML document, elements only, with /*[1]/*[2]/*[2]/*[3] highlighted, its parent marked in purple, and all its ancestors (including the parent) surrounded with a purple circle.

preceding (preceding:: and preceding-sibling::)

Classic tree diagram of generic XML document, elements only, with /*[1]/*[2]/*[2]/*[3] highlighted, its preceding siblings marked in red, and all its preceding elements (including the preceding siblings) surrounded with a red circle.

following (following:: and following-sibling::)

Classic tree diagram of generic XML document, elements only, with /*[1]/*[2]/*[2]/*[3] highlighted, its following sibling marked in blue, and all its following elements (including the following sibling) surrounded with a blue circle.

axes :-)

Combination of above 4

XPath …

… lets us (among other things) select nodes in the tree.

… is used by XSLT, XQuery, Schematron; XLink and XPointer, too.

basic filepath-like path expressions

A bare-bones path expression is similar to filesystem addressing: if the path starts with a solidus (/ aka forward slash), then it represents a path from the root; if it does not start with a solidus then it represents a path from here.

/TEI/teiHeader/fileDesc/titleStmt/title list/item/label

Try it!

In oXygen open the file ham.xml. In the upper L corner there is a text-entry field (looks like a search box). Ensure the box is labelled XPath 2.0 (or XPath 3.0). Then type in /TEI/teiHeader/fileDesc/titleStmt/author. Notice that you have content completion, so it is very easy to type. (If you don’t, set it in Options / Preferences ; Editor / Content Completion / XPath.)

oXygen XPath console An image of an oXygen window

Here is what the oXygen XPath toolbar (or console) will look like.

XPath basics steps are separated by slashes (/, aka solidus) the default axis (i.e., direction) is child:: thus /duck and /child::duck are equivalent to use a different axis, specify it with two colons, e.g. preceding-sibling::item each step selects 0 or more nodes, in this case elements, and hands them on to the next step a step can address attributes, comments, processing instructions, or text nodes, too — we’ll start with elements for now, though the nodes an XPath returns are specified by the last step moose/calf returns all the calf elements that are children of a moose element calf/parent::moose returns all the moose elements that are the parent of a calf element (unlike with real moose, each calf has only one parent moose)

XPath axes: me, my descendants, and my ancestors short long means try it! . self::node() me, whatever I happen to be (element, attribute, comment, processing instruction, document root, or even text) (note: long and short are not exactly equivalent) oXygen shows node cursor is in Note: if you turn on XPath update on cursor move, oXygen will show an XPath to the cursor in the XPath toolbar said child::said my said children in this encoding there is 1 div for each act, each of which holds 1 div for each scene; list all the head elements for the scenes. * child::* all my element children list all children of sourceDesc //div descendant::div my div descendants (note: long and short are not exactly equivalent) list all stage directions (stage) .. parent::node() my parent, whatever it happens to be (element or root) list the parents of stage directions [none] parent::author my parent, if it is an author list the speech (sp) parents of stage directions [none] ancestor::div my div ancestors list all the headings of the div ancestors of the famous To be, or not to be (which you can find by searching for it) [none] ancestor::* all my ancestors list the ancestors of the author element(s)

Kay, chapter 9

XPath axes: following, preceding, and attributes short long means try it! [none] following::item all the items that come after me find all the title elements that occur after the titleStmt [none] following-sibling::l the l children of my parent that come after me (i.e., the rest of the stanza) find all the children of titleStmt that follow author [none] preceding::pb the pbs that come before me find all the p elements that occur before To be, or not to be: [none] preceding-sibling::head the head children of my parent that come before me do any elements ever occur before the speaker within a speech? @rend attribute::rend my rend attribute list all the who attributes @* attribute::* all of my attributes do any of the choice elements have attributes at all?

Kay, chapter 9

Don’t want them all? Many of the above XPaths return multiple nodes — what if you only want a particular one? If you only want Act 3, Scene 1: /TEI/text/body/div[3]/div[1]> Works well presuming you know what you want by element count. But in many cases, that is at least inconvenient, if not outright unknown. No matter how many divs there are, we know this particular scene has an xml:id of sha-ham301. Thus: //div[ @xml:id eq 'sha-ham301'] selects the same node (as would //*[ @xml:id eq 'sha-ham301'])

predicates

An XPath predicate filters the nodes retrieved by a given step

Predicates are expressed after the node test in square brackets

The following are based on ../../demos/xslt_intro/places.xml, which you can open in oXygen with File > Open URL… or ⌘-u (Mac) or ctl-u (Windows), and then pasting in the URL.

XPath selects //listPlace/place[1] the ﬁrst place of each listPlace (of which there only happens to be one) //*[@cRef] all elements that have a cRef attribute //title[ @level eq 'm'] all monographic titles /TEI/text//name[ not( @key ) ] name elements that are missing their key attributes //lg[@type='song']/l[1] list ﬁrst line of each song (16 nodes) (//lg[@type='song']/l)[1] returns ﬁrst line of all songs (1 node)

Repeat Sample Document Instance Warp Speed, Ms Bright! There was a young lady named Bright, Who travelled much faster than light, She departed one day, In a relative way way, And returned on the previous night. See .

XSLT View

Same document, but adding text and attribute nodes

Tree diagram of the limerick Warp Speed, Ms Bright! that has element, attribute, and text nodes except for whitespace-only text nodes.

Element nodes have just the element’s local name; attribute nodes have just the name of the attribute; text nodes have just the word text.