| 摘要 |
A system is described that can aid the querying of structured data stored in a computer system, with particular emphasis on a class of complex queries representing cross tabulations and queries requiring the building of models for their resolution. In the first step the system reads the description of query languages supported by one or more information storage systems; it goes on to read the data dictionaries or other metadata that define the data stored within such systems and finally it optionally reads translation lists concerning the data. The user is then requested to enter a set of keywords describing the search that is needed. The information gathered in the first step is searched and the system returns the most relevant queries as a set of links sorted by relevance. The responses include readable definitions of the queries and more formal query specifications. The system can also display a more accurate representation of the query using associated keywords, thus providing relevance feedback. When a link is selected the underlying query may be run and the results from the data returned. The mechanism is also applicable to formulating complex queries against data models.
|
