<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2 Final//EN">
<HTML>
<HEAD>
<TITLE>Xtract: a query language for XML documents</TITLE>
</HEAD>
<BODY>
<CENTER>
<H1><EM>Xtract</EM>: a query language for XML documents</H1>
</CENTER>
<CENTER>
<EM>Malcolm Wallace, </EM>
<EM>Colin Runciman</EM>
<BR />
University of York
</CENTER>
<CENTER><B>December 1998,<BR />
updated June, August 1999, February 2000</B></CENTER>
<H2>Introduction</H2>
<P> <EM>Xtract</EM>
is a query language based originally on XQL, which was a
W3C proposal that eventually mutated into XPath and XQuery.
The syntax of Xtract is very similar to XPath, although
not completely conformant.
</P>
<P> The idea of
<EM>Xtract</EM>
is that it can be used as a kind of XML-grep at the
command-line, but it could also be used within a scripting
language (such as the Haskell XML combinator library) as a
shorthand for a complicated selection filter.
</P>
<P> All queries return a sequence of XML document fragments
(either whole tagged elements or text contained inside an
element): for our purposes, we also treat attribute
values as document fragments.
</P>
<P>This document describes the expression language for queries.
</P>
<H2>Queries</H2>
<P> Just as in XPath, a query looks rather like a Unix file path,
where the ``directories'' are tags of parent nodes, and the
<TT>/</TT>
separator indicates a parent/child relationship. Hence,
<PRE> matches/match/player </PRE>
selects the
<TT>player</TT>
elements inside
<TT>match</TT>
elements inside a
<TT>matches</TT>
element. The star
<TT>*</TT>
can be used as a wildcard meaning any tagged element, thus:
<PRE> matches/*/player </PRE>
means the
<TT>player</TT>
elements inside
<EM>any</EM>
element within a
<TT>matches</TT>
element. The star can also be used as a suffix or prefix to
match a range of tags:
[
Note that this is
<EM>not</EM>
a full regular expression language: we just provide for the
common cases of wildcards.
]
<PRE> html/h* </PRE>
means all the headings (
<TT><H1></TT>
to
<TT><H6></TT>
) within an HTML document (and HR too!). A double slash
indicates a recursive search for an element tag, so
<PRE> matches//player </PRE>
means all
<TT>player</TT>
elements found at any depth within a
<TT>matches</TT>
element. The plain text enclosed within a tagged element is
expressed with a dash symbol:
<PRE> matches/location/- </PRE>
means the plain text of the location, without any surrounding
<TT><location></TT>
tags. Likewise,
<PRE> *//- </PRE>
simply means to flatten the text of the document at all
levels, removing all tagged element structure. The union of
two queries is expressed with the
<TT>+</TT>
operator and parentheses if required:
<PRE> matches/match/(player + spectators) </PRE>
gives both the players and spectators at a match.
Finally,
<PRE> matches//player/@goals </PRE>
returns the value of the attribute `goals' on the selected
player elements, if the attribute appears.
<H2>Predicates</H2>
There is a notion of a predicate on an element. The square
bracket notation is used:
<PRE> matches/match[player] </PRE>
means all
<TT>match</TT>
elements which contain at least one
<TT>player</TT>
element. It is the match elements that are returned, not
the players they contain. One can also ask for the presence
of a particular attribute:
<PRE> *//player[@goals] </PRE>
means those players (found anywhere within the tree) who
scored any goals. You can compare attribute values using
any of the operators
<TT>=</TT>
,
<TT>!=</TT>
,
<TT><</TT>
,
<TT><=</TT>
,
<TT>></TT>
,
<TT>>=</TT>
all of which use
<EM>lexicographical</EM>
ordering. In this example:
<PRE> */match[player/@surname!=referee/@surname] </PRE>
we only want those matches in which the referee does not
have the same surname as any of the players. A comparison
may be either against another attribute value, or against a
literal string; however a literal string may only appear to
the
<EM>right</EM>
of the operator symbol. For instance,
<PRE> */match[player/@name='colin'] </PRE>
asks for only those matches in which the player named
``colin'' participated. If lexicographical comparison is
inappropriate, numeric comparisons are also possible: these
comparison operators are surrounded by dots:
<TT>.=.</TT>
,
<TT>.!=.</TT>
,
<TT>.<.</TT>
,
<TT>.<=.</TT>
,
<TT>.>.</TT>
,
<TT>.>=.</TT>
Again, either two attribute values are compared, or one
attribute value is compared with a literal integer. For
instance
<PRE> */match[@ourgoals .>. @theirgoals] </PRE>
asks for the matches we won, while
<PRE> */match[@ourgoals .<=. 3] </PRE>
asks for the matches in which we scored three or fewer
goals. (Note that the literal integer is not surrounded by
quote marks.)
<P>
In addition to comparing attribute values, you can also
compare the textual content of elements. For instance,
<PRE> */match[player/- = 'Colin'] </PRE>
asks for the matches in which ``Colin'' participated,
where the name is recorded between the player tags, rather
than as an attribute. All the same conditions and operations
apply as for attribute value comparisons. Note however that
you can only compare texts, not whole structures.
<H2>Combining predicates</H2>
Predicates can be combined using the common Boolean operations
<TT>&</TT>
<TT>|</TT>
and
<TT>~</TT>
, with parentheses for disambiguation if they are required:
<PRE>
*/match/[@ourgoals .=. @theirgoals | (player/@name='colin'
& ~(@opposition='city'))]
</PRE>
means the matches which either ended in a draw, or in which
``colin'' played but the opposition was not ``city''.
<H2>Positional selection</H2>
The final feature of
<EM>Xtract</EM>
is that the square bracket notation is overloaded to allow
the selection of elements by position:
<PRE> */match[3] </PRE>
means the fourth match in the sequence (numbering starts at
zero). You can have a series of indexes, separated by
commas, and ranges are indicated with a dash. The dollar
symbol means the last in the sequence. For example:
<PRE> */match[0,2-$,1] </PRE>
reorders the matches to place the second one last.
<H2>Complex queries</H2>
The full expression language is highly recursive,
permitting you to build arbitrarily complex queries. For
instance:
<PRE> */match[player/@name='colin'][5-$]/referee[@age.>=.34]
</PRE>
means: from the sixth onwards, of those matches in which
``colin'' was a player, select those referees who are older
than 34.
<H2>Grammar</H2>
<P>We give a full grammar for <EM>Xtract</EM>. </P>
<TABLE>
<TR>
<TD><EM>textquery</EM></TD>
<TD>= </TD>
<TD><EM>query</EM></TD>
<TD>elements</TD>
</TR>
<TR><TD /><TD>| </TD><TD><TT>-</TT></TD><TD>plain text</TD></TR>
<TR></TR>
<TR>
<TD><EM>query</EM></TD>
<TD>= </TD>
<TD><EM>string</EM></TD>
<TD>tagged element</TD>
</TR>
<TR>
<TD />
<TD>| </TD>
<TD><EM>string</EM><TT>*</TT></TD>
<TD>prefix of tag</TD>
</TR>
<TR>
<TD />
<TD>| </TD>
<TD><TT>*</TT><EM>string</EM></TD>
<TD>suffix of tag</TD>
</TR>
<TR><TD /><TD>| </TD><TD><TT>*</TT></TD><TD>any element</TD></TR>
<TR>
<TD />
<TD>| </TD>
<TD><TT>( </TT><EM>textquery </EM><TT>)</TT></TD>
<TD>grouping</TD>
</TR>
<TR>
<TD />
<TD>| </TD>
<TD><EM>query</EM><TT>/</TT><EM>textquery</EM></TD>
<TD>direct descendant</TD>
</TR>
<TR>
<TD />
<TD>| </TD>
<TD><EM>query</EM><TT>//</TT><EM>textquery</EM></TD>
<TD>deep descendant</TD>
</TR>
<TR>
<TD />
<TD>| </TD>
<TD><EM>query</EM><TT>/@</TT><EM>string</EM></TD>
<TD>value of attribute</TD>
</TR>
<TR>
<TD />
<TD>| </TD>
<TD><EM>query </EM><TT>+ </TT><EM>textquery</EM></TD>
<TD>union</TD>
</TR>
<TR>
<TD />
<TD>| </TD>
<TD><EM>query</EM><TT>[</TT><EM>predicate</EM><TT>] </TT></TD>
<TD>predicates</TD>
</TR>
<TR>
<TD />
<TD>| </TD>
<TD><EM>query</EM><TT>[</TT><EM>positions</EM><TT>] </TT></TD>
<TD>indexing</TD>
</TR>
<TR></TR>
<TR>
<TD><EM>qa</EM></TD>
<TD>= </TD>
<TD><EM>textquery</EM></TD>
<TD>has tagged element</TD>
</TR>
<TR>
<TD /><TD>| </TD><TD><EM>attribute</EM></TD><TD>has attribute</TD>
</TR>
<TR></TR>
<TR>
<TD><EM>predicate</EM></TD>
<TD>= </TD>
<TD><EM>qa</EM></TD>
<TD>has tagged element or attribute</TD>
</TR>
<TR>
<TD />
<TD>| </TD>
<TD><EM>qa </EM><EM>op </EM><EM>qa</EM></TD>
<TD>lexical comparison of attribute values or element texts</TD>
</TR>
<TR>
<TD />
<TD>| </TD>
<TD>
<EM>qa </EM><EM>op </EM><TT>'</TT><EM>string</EM><TT>'</TT>
</TD>
<TD>lexical comparison of attribute value or element text</TD>
</TR>
<TR>
<TD />
<TD>| </TD>
<TD>
<EM>qa </EM><EM>op </EM><TT>"</TT><EM>string</EM><TT>"</TT>
</TD>
<TD>lexical comparison of attribute value or element text </TD>
</TR>
<TR>
<TD />
<TD>| </TD>
<TD><EM>qa </EM><EM>nop </EM><EM>qa</EM></TD>
<TD>numeric comparison of attribute values or element texts</TD>
</TR>
<TR>
<TD />
<TD>| </TD>
<TD><EM>qa </EM><EM>nop </EM><EM>integer</EM></TD>
<TD>numeric comparison of attribute value or element text</TD>
</TR>
<TR>
<TD />
<TD>| </TD>
<TD><TT>( </TT><EM>predicate </EM><TT>)</TT></TD>
<TD>grouping</TD>
</TR>
<TR>
<TD />
<TD>| </TD>
<TD><EM>predicate </EM><TT>& </TT><EM>predicate</EM></TD>
<TD>logical and</TD>
</TR>
<TR>
<TD />
<TD>| </TD>
<TD><EM>predicate </EM><TT>| </TT><EM>predicate</EM></TD>
<TD>logical or</TD>
</TR>
<TR>
<TD />
<TD>| </TD>
<TD><TT>~ </TT><EM>predicate</EM></TD>
<TD>logical not</TD>
</TR>
<TR></TR>
<TR>
<TD><EM>attribute</EM></TD>
<TD>= </TD>
<TD><TT>@</TT><EM>string</EM></TD>
<TD>attribute of this element</TD>
</TR>
<TR>
<TD />
<TD>| </TD>
<TD><EM>query</EM><TT>/@</TT><EM>string</EM></TD>
<TD>attribute of descendant</TD>
</TR>
<TR></TR>
<TR>
<TD><EM>positions</EM></TD>
<TD>= </TD>
<TD><EM>position </EM>{ <TT>, </TT><EM>positions</EM>} </TD>
<TD>comma-separated sequence</TD>
</TR>
<TR>
<TD />
<TD>| </TD>
<TD><EM>position </EM><TT>- </TT><EM>position</EM></TD>
<TD>range</TD>
</TR>
<TR></TR>
<TR>
<TD><EM>position</EM></TD>
<TD>= </TD>
<TD><EM>integer</EM></TD>
<TD>positions start at zero</TD>
</TR>
<TR><TD /><TD>| </TD><TD><TT>$</TT></TD><TD>last element</TD></TR>
<TR></TR>
<TR>
<TD><EM>op</EM></TD>
<TD>= </TD>
<TD><TT>=</TT></TD>
<TD>lexical equality</TD>
</TR>
<TR>
<TD /><TD>| </TD><TD><TT>!=</TT></TD><TD>lexical inequality</TD>
</TR>
<TR>
<TD /><TD>| </TD><TD><TT><</TT></TD><TD>lexically less than</TD>
</TR>
<TR>
<TD />
<TD>| </TD>
<TD><TT><=</TT></TD>
<TD>lexically less than or equal</TD>
</TR>
<TR>
<TD />
<TD>| </TD>
<TD><TT>></TT></TD>
<TD>lexically greater than</TD>
</TR>
<TR>
<TD />
<TD>| </TD>
<TD><TT>>=</TT></TD>
<TD>lexically greater than or equal</TD>
</TR>
<TR></TR>
<TR>
<TD><EM>nop</EM></TD>
<TD>= </TD>
<TD><TT>.=.</TT></TD>
<TD>numeric equality</TD>
</TR>
<TR>
<TD /><TD>| </TD><TD><TT>.!=.</TT></TD><TD>numeric inequality</TD>
</TR>
<TR>
<TD /><TD>| </TD><TD><TT>.<.</TT></TD><TD>numeric less than</TD>
</TR>
<TR>
<TD />
<TD>| </TD>
<TD><TT>.<=.</TT></TD>
<TD>numeric less than or equal</TD>
</TR>
<TR>
<TD />
<TD>| </TD>
<TD><TT>.>.</TT></TD>
<TD>numeric greater than</TD>
</TR>
<TR>
<TD />
<TD>| </TD>
<TD><TT>.>=.</TT></TD>
<TD>numeric greater than or equal</TD>
</TR>
</TABLE>
</BODY>
</HTML>
|
