Oracle International Corporation

Anjana Manian

K Karun

Ravi Murthy

Eric Sedlar

Sivasankaran Chandrasekar

Olga Peschansky

Dmitry Lenkov

Kwok Lun Alex Yiu

A compact binary encoding technique for information that is logically hierarchically structured, such as XML data, maintains all of the features of XML data in a useable form, such as the hierarchical structure underlying the data. Hence, data encoded in this format can undergo XML-based processing on-the-fly as it is being received or fetched, as if the data was being processed linearly in its textual character-based format. Processing of data encoded in this format can begin without having to wait for and decode the entire data set. The overhead due to XML tags is significantly minimized. The encoded data can be processed more efficiently because the data is pre-parsed. Values may be stored in their native type formats and, therefore, processing of the encoded data avoids costly type conversions. Further, any available structural constraint information can be effectively exploited.

Encoding of hierarchically organized data for efficient storage and processing

Described herein are approaches that allow an XML entity to be accessed in a way that requires less memory. These approaches involve dynamically generating and maintaining an in-memory representation of only a portion of an XML tree. The in-memory representation of an XML tree is herein referred to as a node tree. The node tree contains data from the XML tree, and is generated by extracting data from a compressed form of an XML entity. In addition, the node tree contains information about the location of specific elements within the compressed XML entity. The approaches described herein allow an XML tree to be accessed without having to generate an in-memory representation of the whole XML tree, thus reducing the amount of memory needed to access the data in the XML tree.

Reducing memory requirements needed to represent XML entities

Growing community of inventors

Newark, CA, United States of America

Anjana Manian