The W3C XQuery Working Group has updated XQuery 1.0 and XPath 2.0 Full-Text Use Cases and posted the first public working draft of XQuery 1.0 and XPath 2.0 Full-Text. According to the latter,
XML documents may contain highly-structured data (numbers, dates), unstructured data (untagged free-flowing text), and semi-structured data (text with embedded tags). Where a document contains unstructured or semi-structured data, it is important to be able to search that data using Information Retrieval techniques such as full-text search. Full-text search is different from substring search in many ways:
A full-text search searches for phrases (a sequence of words) rather than substrings. A substring search for news items that contain the string "lease" will return a news item that contains "Foobar Corporation releases the 20.9 version ...". A full-text search for the phrase "lease" will not.
There is an expectation that a full-text search will support language- and token-based searches which substring search cannot. An example of a language-based search is "find me all the news items that contain a word with the same linguistic stem as "mouse" (finds "mouse" and "mice"). An example of a token-based search is "find me all the news items that contain the word "XML" within 3 words (tokens) of "Query".
Full-text search is subject to the vageries and nuances of language. The results it returns are often of varying usefulness. When you search a web site for all cameras that cost less than $100, this is an exact search. There is a set of cameras that match this search, and a set that do not. Similarly, when you do a string search across news items for "mouse", there is only 1 expected result set. When you do a full-text search for, say, all the news items that contain the word "mouse", you probably expect to find news items with the word "mice", and possibly "rodents" (or possibly "computers"!). But not all results are equal : some results are more "mousey" than others. Because full-text search can be inexact, we have the notion of score or relevance : we generally expect to see the most relevant results at the top of the results list. Of course, relevance is in the eye of the beholder. Note: as XQuery/XPath evolves, it may apply the notion of score to querying structured search. For example, when making travel plans or shopping for cameras, it is sometimes more useful to get an ordered list of near-matches. If XQuery/XPath defines a generalized inexact match, we assume that XQuery/XPath can utilize the scoring framework provided by the full-text language.
As XML becomes mainstream, users expect to be able to store and search all their documents in XML. This requires a standard way to do full-text search, as well as structured searches, against XML documents. A similar requirement for full-text search led ISO to define the SQL/MM-FT standard. SQL/MM-FT defines extensions to SQL to express full-text queries providing similar functionality as this full-text language extension to XQuery 1.0/XPath 2.0 does.
Full-text queries are performed on text which has been tokenized, i.e., broken into a sequence of words, units of punctuation, and spaces.
A word is defined as any character, n-gram, or sequence of characters returned by a tokenizer as a basic unit to be queried. Each instance of a word consists of one or more consecutive characters. Beyond that, words are implementation defined. Note that consecutive words need not be separated by either punctuation or space, and words may overlap. A phrase is a sequence of ordered words which can contain any number of words.
Tokenization enables functions and operators which work with the relative positions of words (e.g., proximity operators). It also uniquely identifies sentences and paragraphs in which words appear. Tokenization also enables functions and operators which operate on a part or the root of the word (e.g., wildcards, stemming).
We use the namespace "ft" (for full-text) that corresponds to the URL http://www.w3.org/2004/07/xquery-full-text and defines the namespace of full-text search. We also use "fts" for definitional purposes in semantics Section.