Semi-structured data storage schema selection
AssigneeMicrosoft Corporation
InventorsJi-Rong Wen, Shihui Zheng, Hongjun Lu
Life of this patent
Estimating the term…
Abstract
In one aspect, this disclosure relates to a method and associated apparatus that allows a user to obtain a semi-structured data input and a workload input. An improved semi-structured data storage schema is selected for a relational schema in response to the semi-structured data input and the workload input. The semi-structured data is segmented based on the selected improved semi-structured data storage schema. In one aspect, the semi-structured data is XML data.
Claims80
A method comprising:
- obtaining semi-structured data;
- obtaining workload input;
- selecting an improved semi-structured data storage schema for a relational schema in response to the semi-structured data and the workload input;
- and
- segmenting the semi-structured data based on the selected improved semi-structured data storage schema.
An apparatus comprising:
- an XML data storage schema portion for obtaining XML data;
- a workload schema portion for obtaining workload input;
- and
- an XML storage schema selection portion for selecting an improved relational schema in response to the XML data and in response to the workload input, wherein the XML data storage schema selection portion segments the XML data based on the selected improved XML data storage schema.
An apparatus comprising:
- a computer environment including a computer processor, the computer processor including:
- a semi-structured data storage schema portion for obtaining semi-structured data,
- a workload schema portion for obtaining workload input,
- a semi-structured data storage schema selection portion for selecting an improved semi-structured data storage schema for a relational schema in response to the semi-structured data and in response to the workload input, wherein the semi-structured data storage schema selection portion segments the semi-structured data based on the selected improved semi-structured data storage schema, wherein the semi-structured data storage schema selection portion is based on a hillclimbing algorithm, and
- a relational database for storing the semi-structured data storage schema in the relational schema.
A computer readable medium having computer executable instructions for:
- obtaining XML data including an initial state;
- obtaining workload input;
- selecting an improved XML data storage schema for a relational schema in response to the XML data and in response to the workload input;
- segmenting the XML data based on the selected improved XML data storage schema;
- and
- transforming the initial state to a neighboring state.
An H-Cut method for transforming a semi-structured data storage schema into a configuration, the method comprising:
- splitting a first instance that is related to a second instance and a third instance into a first split instance that is related to the second instance and a second split instance that is related to the third instance.
An H-Merge method for transforming a semi-structured data storage schema into a configuration, the method comprising:
- merging a first instance that is related to a second instance and a third instance that is related to a fourth instance into a fifth instance that is related to the second instance and the fourth instance.
A method, comprising:
- applying a cut operation with a first operand;
- applying a merge operation with a second operand, wherein a plurality of possible resulting neighboring states can result;
- running the hillclimbing algorithm for the possible resulting neighboring states;
- determining whether a transformation will reduce or increase the cost of the workload;
- ignoring those transformations that increase the cost of the workload;
- and
- considering those transformations that decrease the cost of the workload.
An apparatus comprising:
- a semi-structured data storage schema portion for obtaining semi-structured data;
- a workload schema portion for obtaining workload input;
- and
- a semi-structured data storage schema selection portion for selecting an improved relational schema in response to the semi-structured data and in response to the workload input, wherein the semi-structured data storage schema selection portion segments the semi-structured data based on the selected improved semi-structured data storage schema.
An H-Cut method for transforming an XML data storage schema into a configuration, the method comprising:
- splitting a first instance that is related to a second instance and a third instance into a first split instance that is related to the second instance and a second split instance that is related to the third instance.
Dependencies are inferred from the claim text.