Optimizing differential XML processing by leveraging schema and statistics
View Patent ↗Statistical information about instance documents and schema information are used to integrate multiple state transitions that enable sectioning of a structure document, thereby generating an optimum automaton. In integrating state transitions, consecutively matching state transitions are held in the form of an ID list, which is then used to count the number of consecutive state transitions. Furthermore, patterns in the number of occurrences of repetitive elements including nested elements are statistically obtained. Variations of blanks in XML are addressed by using a statistical method. Schema information is used to build an automaton beforehand, thereby initialization overhead of the syntax parsing apparatus is reduced.
1. A structured document processing method for performing syntax parsing of a structured document in the form of electronic data, comprising:
generating a state transition sequence of a plurality of states enabling sectioning of a structured document into a plurality of nodes;
integrating state transitions in the state transition sequence generated at the automaton generating by using statistical information regarding an instance document which is the entity of the structured document and statistically obtaining patterns in the number of occurrences of repetitive elements in the state transitions by using the statistical information;
integrating the state transitions in the state transition sequence generated at the automaton generating by using schema information which defines the structure and format of information regarding the structured document; and
mutually optimizing automatons integrated at the instance document analyzing and the schema information analyzing,
wherein the structured document comprises an XML document,
wherein the plurality of states enabling sectioning are defined by SAX events, and
wherein said method further comprises assigning an ID to each of the state transitions in integration of the multiple state transitions at the statistically obtaining, storing consecutively matching state transitions in the form of a list of IDs, and counting the occurrences of the consecutively matching state transitions by using the list of IDs,
wherein the automaton optimizing optimizes the repetitive elements even if the repetitive elements are nested, and
wherein a pattern of any number of blank characters appearing between elements in the structured document is fixed by using the statistical information at the statistically obtaining.