Database Reference
In-Depth Information
TraCX: Transformation of Compressed XML
Stefan Böttcher, Rita Hartel, and Sebastian Stey
University of Paderborn, Computer Science, Fürstenallee 11, 33102 Paderborn, Germany
{stb@,rst@,sebstey@mail.}uni-paderborn.de
Abstract.
While the volume of XML data and sometimes even the processing
time of XML data can be reduced by using XML compression and storing,
processing, and transferring compressed XML instead of uncompressed XML, a
transformation of the transferred XML data into the receiver's XML format via
XQuery cannot be performed on the compressed XML data directly. Instead,
XQuery transformation of compressed XML data requires a prior decompres-
sion. In this paper, we present a generic approach to transform compressed or
uncompressed XML representations that support basic navigation and update as
well as optional copy functionalities. In a series of experiments, we have shown
that using our approach to transform compressed XML is not only faster than
the indirect approach via decompression, XQuery transformation, and recom-
pression, but also that our approach transforms compressed XML as efficient as
other XQuery evaluators transform uncompressed XML only.
Keywords:
XML compression, XQuery, XML transformation.
1 Introduction
1.1 Motivation
Due to its flexible structure, XML enjoys increasing popularity as a data exchange
format and as an intermediate storage format, e.g. in production chains where XML
data is exchanged, transformed and eventually further processed sequentially by mul-
tiple participants. But the overhead caused by this flexible structure forms the biggest
disadvantage of using XML as a data exchange or storage format. Transferring and
storing a compressed XML representation instead of uncompressed XML might help
to solve this problem. However, when XML files have to be shared with other parties,
often the XML structure needs to be adapted to match the XML data structure on the
receiver's side. A typical approach to adapt uncompressed XML files is to transform
them via XQuery [1] into the format that is required on the receiver's side. But when
exchanging compressed XML, the decompression prior to the XQuery transformation
and, after the XQuery transformation, the recompression of transformed XML data
needed for compact storage and for further transferal of compressed XML data may
be a significant overhead. Instead, the XQuery transformation of compressed data
appears to be a promising alternative. However, up to now, XQuery transformation of
compressed XML data has been very limited, i.e., up to now, there has not been any
generic approach to XQuery evaluation applicable to all XML compression tech-
niques. Instead, most XML compression techniques do not support XQuery at all, and
