Unlock instant, AI-driven research and patent intelligence for your innovation.

Partition filtering using in-memory smart indexes

A storage and indexing technology, applied in database indexing, instrumentation, structured data retrieval, etc.

Active Publication Date: 2020-06-19
MICROSOFT TECH LICENSING LLC
View PDF3 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Table-entity-attribute organization for partitions with partition keys can optimize some operations, but can face challenges in supporting others
For example, it can be challenging to filter on specific fields (e.g., attribute names) rather than partition keys or partitions on row keys

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Partition filtering using in-memory smart indexes
  • Partition filtering using in-memory smart indexes
  • Partition filtering using in-memory smart indexes

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0016] Example apparatus and methods provide different schemes for filtering partitions on specific fields when the partitions are associated with a table-entity-attribute organization. Partitions can be initially stored in a table storage system. A table storage system can include one or more data warehouses. Partitions can be moved to storage that supports one or more indexes that can be used to filter partitions. The memory may eg be a shared memory or a cache. The storage may be a data warehouse located external to the table storage system.

[0017] Table storage systems can use table-entity-attribute organization to store large amounts of structured but non-relational data. Data can be stored in tables using, for example, binary large objects (BLOBs). Tables can be clusters of entities. Entities in a table may roughly correspond to database records. An attribute in an entity can roughly correspond to a field in a database record, where a field can have a name and a ...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

Example apparatus and methods detect queries to partitions in a table store arranged in a table-entity-attribute organization. A partition is a cluster of entities with the same partition key value. If a query seeks to filter on specific entities in a partition, the partition can be copied to memory (eg, structured shared cache, shared memory) prepared with one or more indexes that facilitate filtering on the partition. Example apparatus and methods may copy partitions from a table store into storage and then use one or more indexes to filter the partitions on specific fields from storage. Filtering partitions on a particular field can be done in parallel, possibly massively in parallel, using all indexes available in memory.

Description

Background technique [0001] Structured Query Language (SQL) database models rely on tabular relationships in relational databases. NoSQL, which stands for "not only" SQL, refers to schemes for storing and retrieving data that do not use the SQL model. NoSQL models can use key-value pairs in entities in table storage optimized for retrieval and appending. Optimization for one action may cause other actions to have suboptimal performance. A key-value store allows data to be stored in a schemaless manner, such as objects (for example, binary large objects (BLOBs)). Table storage systems can be built on BLOB and NoSQL frameworks. Like all data structures and organizations, a table storage system may work well for one type of operation, but may not work well for another type of operation. [0002] Table storage systems can store large amounts of data using NoSQL databases. The data may eg be unstructured data. The data can also be, for example, structured but non-relational d...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Patents(China)
IPC IPC(8): G06F16/22G06F16/23G06F16/2453
CPCG06F16/22G06F16/2453G06F16/23G06F16/2228
Inventor R.凯姆K.茹安尼F.吉拉尔W.茹安尼
Owner MICROSOFT TECH LICENSING LLC