Query access optimizations for tiered index architecture
AssigneeSybase, Inc.
InventorsAnant Agarwal, Steven A. Kirk, Blaine French, Nandan Marathe, Shailesh Mungikar, Kaushal Mittal
Life of this patent
Estimating the term…
Abstract
Disclosed herein are methods for retrieving data from a database. An embodiment operates searching for a key in a first index. The method determines that the searching will require a storage access request and issues the storage access request. The method continues searching for the key in a second index.
Claims20
A database system for managing index data density in a tiered indexed architecture, the database system comprising:
- a memory;
- and
- at least one processor, coupled to the memory, and configured to perform operations comprising:
- creating a selection tree, the selection tree comprising a set of keys representing contents of a group of sorted data sets;
- determining a first champion key from the set of keys;
- determining a second-place key from the set of keys, wherein to determine the second-place key, the at least one processor is further configured to perform operations comprising:
- recording a number of times the candidate key has been equal to the second-place key out of a number of retrieval operations;
- determining the second-place key when the number exceeds a threshold;
- and
- omitting the determining the second-place key when the number does not exceed the threshold;
- removing, in response to a data request, the first champion key from the selection tree;
- comparing a candidate key with the second-place key;
- and
- determining a second champion key.
A computer-implemented method for retrieving data from a database having a tiered index architecture balancing index data density, the computer-implemented method comprising:
- creating, by at least one processor, a selection tree, the selection tree comprising a set of keys representing contents of a group of sorted data sets;
- determining, by the at least one processor, a first champion key from the set of keys;
- determining, by the at least one processor, a second-place key from the set of keys, the determining the second-place key further comprising:
- recording, by the at least one processor, a number of times the candidate key has been equal to the second-place key out of a number of retrieval operations;
- determining, by the at least one processor, the second-place key when the number exceeds a threshold;
- and
- omitting, by the at least one processor, the determining the second-place key when the number does not exceed the threshold;
- removing, by the at least one processor, in response to a data request, the first champion key from the selection tree;
- comparing, by the at least one processor, a candidate key with the second-place key;
- and
- determining, by the at least one processor, a second champion key.
A non-transitory computer-readable storage medium having computer-executable instructions stored thereon, wherein the instructions, when executed, cause at least one computer processor to perform operations on a database having a tiered index architecture balancing index data density, the operations comprising:
- creating a selection tree, the selection tree comprising a set of keys representing contents of a group of sorted data sets;
- determining a first champion key from the set of keys;
- determining a second-place key from the set of keys, wherein to determine the second-place key, the operations further comprise:
- recording a number of times the candidate key has been equal to the second-place key out of a number of retrieval operations;
- determining the second-place key when the number exceeds a threshold;
- and
- omitting the determining the second-place key when the number does not exceed the threshold;
- removing, in response to a data request, the first champion key from the selection tree;
- comparing a candidate key with the second-place key;
- and
- determining a second champion key.
Dependencies are inferred from the claim text.