Efficient way to evaluate uncorrelated path-based row sources with XML storage
US8239373B2 · kind B2 · utility
Assignee
Inventors
Key dates
| Filing date | May 18, 2009 |
| Grant date | Aug 7, 2012 |
| Priority date | — |
| Expiry date | Feb 11, 2030 |
Classification
- Technology area (CPC G)Physics
- CPC primaryG06F16/8358
- WIPO fieldComputer technology
- WIPO sectorElectrical engineering
Abstract
The approaches described herein provide an efficient way to process certain kinds of queries that retrieve XML data stored in an object-relational database. A technique determines whether a query comprises a plurality of iteration expressions that exhibit a particular relationship to one another, and if so, re-writes the query and produces an optimized query execution plan. The query execution plan specifies an XPath-based row source with caching enabled. A very small cache structure is used, and reference locators are stored in the cache that point into binary-encoded data rather than copying XMLType data into the cache. Row data may be cached on the first use of the path-based row source and re-used in subsequent requests for rows. Once there is a subsequent request for rows from the optimized path-based row source, the cache may be reorganized to efficiently produce the rows of the row source.
Source: USPTO / EPO open patent data. Objective bibliographic and citation counts.