Predictive database index modification

Patent metadata

Field	Value
Publication number	US-11144538-B2
Application number	US-201916412637-A
Country	US
Kind code	B2
Filing date	May 15, 2019
Priority date	May 15, 2019
Publication date	Oct 12, 2021
Grant date	Oct 12, 2021

Abstract

Methods, computer program products, and systems are presented. The method computer program products, and systems can include, for instance: examining data of breaches of a geofence by client computer devices to determine respective positions of the breaches; receiving by a database system transaction data from client computer devices, the transaction data defined by insert query data; changing a structure of an index of a certain table of the database system in response to the receiving; generating logging data in response to the receiving of the transaction data; providing using the logging data prediction specifying data, the prediction specifying data specifying one or more change predicted to occur with respect to the index; and modifying the index in dependence on the prediction specifying data.

First claim

What is claimed is:

1. A computer implemented method comprising:
receiving by a database system transaction data from client computer devices, the transaction data defined by insert query data;
changing a structure of an index of a certain table of the database system in response to the receiving;
generating logging data in response to the receiving of the transaction data;
providing using the logging data prediction specifying data, the prediction specifying data specifying one or more change predicted to occur with respect to the index; and
modifying the index in dependence on the prediction specifying data.

2. The computer implemented method of claim 1 , wherein the prediction specifying data specifies a classifier of an index change pattern, and wherein the modifying includes proactively modifying the index using the classifier of the index change pattern.

3. The computer implemented method of claim 1 , wherein the prediction specifying data specifies a count of leaf pages predicted to be added, and wherein the modifying includes proactively initiating creation of one or more leaf page for the index according to the count.

4. The computer implemented method of claim 1 , wherein the modifying includes proactively allocating one or more leaf page for accommodation of page split data of a certain page of the index.

5. The computer implemented method of claim 1 , wherein the modifying includes proactively initiating creating one or more leaf page for accommodation of page split data of one or more certain leaf page of the index, wherein the proactively initiating creating includes asynchronously initiating creating the one or more leaf page prior a determination that free space of one or more certain leaf page of the index is insufficient to store incoming transaction data.

6. The computer implemented method of claim 1 , wherein the modifying includes proactively initiating creating one or more leaf page for accommodation of page split data of one or more certain leaf page of the index, wherein the proactively initiating creating includes asynchronously initiating creating the one or more leaf page prior a determination that free space of one or more certain leaf page of the index is insufficient to store incoming transaction data, the proactively initiating creating including performing, prior to the determination that free space of one or more certain leaf page of the index is insufficient to store incoming transaction data, one or more of (a) allocating the one or more leaf page to the index from a page store, (b) linking the one or more leaf page to related leaf pages of the index, and (c) copying page split data from an existing one or more leaf page of the index to the one or more leaf page.

7. The computer implemented method of claim 1 , wherein the prediction specifying data specifies that more than one leaf page for the index will be created, and wherein the modifying includes synchronously initiating creating of first and second leaf pages in response to a determination that a certain leaf page of the index has insufficient free space to store incoming transaction data.

8. The computer implemented method of claim 1 , wherein the providing the prediction specifying data includes querying a predictive model that has been subject to training by machine learning using the logging data.

9. The computer implemented method of claim 1 , wherein the providing the prediction specifying data includes querying a predictive model that has been subject to training by machine learning using the logging data, wherein training of the predictive model includes applying for successive time periods a training data set, the training data set including for respective time periods transaction data, index change data, a table ID, and an index change pattern classifier.

10. The computer implemented method of claim 1 , wherein the prediction specifying data specifies a first index change classifier for the index, and a second index change classifier for a second index, wherein the modifying includes proactively modifying the index using the first index change classifier for the index, and proactively modifying the second index using the second index change classifier, wherein proactively modifying the index includes initiating creating a leaf page for the index appended at an end of index after a current end leaf page of the index, wherein proactively modifying the second index includes initiating creating a leaf page for the index added intermediate of a first leaf page and a last leaf page of the second index, wherein the first index change classifier is a classifier specifying that an index change pattern of the first index is an appended leaf page index change pattern, wherein the second index change classifier is a classifier specifying that an index change pattern of the second index is an distributed leaf page index change pattern.

11. The computer implemented method of claim 1 , wherein the generating logging data in response to the receiving of the transaction data includes generating index logging data that specifies the generation of new indexes and new index pages, and wherein the providing using the logging data prediction specifying data includes providing using the index logging data the prediction specifying data.

12. The computer implemented method of claim 1 , wherein the prediction specifying data specifies the prediction that a certain count of intermediate pages and that a particular count of leaf pages will be added to the index.

13. A computer program product comprising:
a computer readable storage medium readable by one or more processing circuit and storing instructions for execution by one or more processor for performing a method comprising:
receiving by a database system transaction data from client computer devices, the transaction data defined by insert query data;
changing a structure of an index of a certain table of the database system in response to the receiving;
generating index logging data in response to the receiving of the transaction data;
providing using the index logging data prediction specifying data, the prediction specifying data specifying one or more page generation change predicted to occur with respect to the index; and
modifying the index in dependence on the prediction specifying data.

14. The computer program product of claim 13 , wherein the prediction specifying data specifies a classifier of an index change pattern, and wherein the modifying includes proactively modifying the index using the classifier of the index change pattern.

15. The computer program product of claim 13 , wherein the prediction specifying data specifies a count of leaf pages predicted to be added, and wherein the modifying includes proactively initiating creation of one or more leaf page for the index according to the count.

16. The computer program product of claim 13 , wherein the modifying includes proactively allocating one or more leaf page for accommodation of page split data of a certain page of the index.

17. The computer program product of claim 13 , wherein the modifying includes proactively initiating creating one or more leaf page for accommodation of page split data of one or more certain leaf page of the index, wherein the proactively initiating creating includes asynchronously initiating creating the one or more leaf page prior a determination that free space of one or more certain leaf page of the index is insufficient to store incoming transaction data.

18. The computer program product of claim 13 , wherein the modifying includes pr

Assignees

• IBM

Inventors

• Li Shuo
• Wang Xiaobo
• Liang Ping
• Li Min
• Wang Ping
• Josten Jeffrey William

Classifications

• G06N7/01

  Probabilistic graphical models, e.g. probabilistic networks · CPC title

• G06F16/2379 Primary

  Updates performed during online database operations; commit processing · CPC title

• G06N5/04

  Inference or reasoning models · CPC title

• G06F16/2272 Primary

  Management thereof · CPC title

• G06N3/006

  based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO] · CPC title