Method and a consistency checker for finding data inconsistencies in a data repository

Patent metadata

Field	Value
Publication number	US-9454561-B2
Application number	US-201214407482-A
Country	US
Kind code	B2
Filing date	Jun 15, 2012
Priority date	Jun 15, 2012
Publication date	Sep 27, 2016
Grant date	Sep 27, 2016

Abstract

A consistency checker and ( 200 ) a method performed by a consistency checker ( 200 ) for finding inconsistencies of data posts stored in a data repository ( 220 ) is provided. The consistency checker ( 200 ) has a storage unit ( 202 ) in which a number of inconsistency patterns are stored, each inconsistency pattern being associated with an inconsistency analysis rule for a value of a data attribute. The consistency checker further has a receiving unit ( 204 ) for receiving a data post from the data repository, and a logic unit ( 206 ) arranged for matching a value of at least one data attribute of the data post with a value of the corresponding data attribute of the stored inconsistency patterns, and if there is a match, applying the inconsistency analysis rule that is connected with the matched inconsistency pattern to the data post to evaluate whether there is an inconsistency in the data post or not, wherein the number of inconsistency patterns each defines an inconsistency that has previously occurred for a value of a data attribute in the data repository.

First claim

The invention claimed is:

1. A method performed by a consistency checker for finding inconsistencies of data posts stored in a data repository, the consistency checker having access to a number of inconsistency patterns, each inconsistency pattern being associated with an inconsistency analysis rule for a value of a data attribute, the method comprising:
receiving a data post from the data repository;
matching a value of a data attribute of the data post with a value of a corresponding data attribute of the stored inconsistency patterns;
if there is a match, applying the inconsistency analysis rule that is connected with the matched inconsistency pattern to the data post to evaluate whether there is an inconsistency in the data post or not, wherein the number of inconsistency patterns each defines an inconsistency that has previously occurred for the value of the data attribute in the data repository,
wherein each of the number of inconsistency patterns has a probability weight, the probability weight indicating a frequency of inconsistency of each of the inconsistency patterns that occurred for the data posts in the data repository; and
if there is more than one match, selecting inconsistency pattern in an inconsistency pattern order according to the probability weight for each matched inconsistency pattern.

2. The method of claim 1 , further comprising:
if the applied inconsistency analysis rule showed that there is an inconsistency for the data post, increasing the probability weight of the inconsistency pattern connected to the applied inconsistency analysis rule.

3. The method of claim 1 , further comprising:
if no match is found, applying another inconsistency analysis rule.

4. The method of claim 3 , further comprising, if the applied another inconsistency analysis rule showed that there is an inconsistency, storing a new inconsistency pattern together with the applied another inconsistency analysis rule, the new inconsistency pattern being associated with a value of at least one data attribute of the data post.

5. The method of claim 1 , further comprising repeating the method for a plurality of data posts in the data repository.

6. The method of claim 1 , further comprising:
removing the inconsistency pattern of the number of inconsistency patterns that has found the least number of inconsistencies.

7. The method of claim 1 , wherein if there is more than one match, the inconsistency pattern with highest probability weight is selected first.

8. A consistency checker for finding inconsistencies of data posts stored in at least one data repository, the consistency checker comprising:
a memory circuit in which a number of inconsistency patterns are stored, each inconsistency pattern being associated with an inconsistency analysis rule for a value of a data attribute;
a receiving circuit for receiving a data post from the data repository; and
a logic circuit arranged for:
matching a value of at least one data attribute of the data post with a value of a corresponding data attribute of the stored inconsistency patterns;
if there is a match, applying the inconsistency analysis rule that is connected with the matched inconsistency pattern to the data post to evaluate whether there is an inconsistency in the data post or not, wherein the number of inconsistency patterns each defines an inconsistency that has previously occurred for the value of the data attribute in the data repository,
wherein each of the number of inconsistency patterns has a probability weight, the probability weight indicating a frequency of inconsistency of each of the inconsistency patterns that occurred for the data posts in the data repository; and
if there is more than one match, selecting inconsistency pattern in an inconsistency pattern order according to the probability weight for each matched inconsistency pattern.

9. The consistency checker of claim 8 , wherein the logic circuit is further arranged for:
if the applied inconsistency analysis rule showed that there is an inconsistency for the data post, increasing the probability weight of the inconsistency pattern connected to the applied inconsistency analysis rule.

10. The consistency checker of claim 8 , the logic circuit being further arranged for, if no match is found, applying another inconsistency analysis rule.

11. The consistency checker of claim 10 , the logic circuit being further arranged for, if the applied another inconsistency analysis rule showed that there is an inconsistency, storing a new inconsistency pattern together with the applied another inconsistency analysis rule, the new inconsistency pattern being associated with a certain value of a data attribute of the data post.

12. The consistency checker of claim 8 , the logic circuit being further arranged for removing the inconsistency pattern of the number of inconsistency patterns stored in a storage unit that has found the least number of inconsistencies.

13. The consistency checker of claim 8 , wherein if there is more than one match, the inconsistency pattern with highest probability weight is selected first.

14. A non-transitory computer-readable medium comprising, stored thereupon, computer program code that, when run in a computer having access to a number of inconsistency patterns, each inconsistency pattern being associated with an inconsistency analysis rule for a value of a data attribute, causes the computer to find inconsistencies of data posts stored in a data repository by:
receiving a data post from the data repository;
matching a value of a data attribute of the data post with a value of a corresponding data attribute of the stored inconsistency patterns;
if there is a match, applying the inconsistency analysis rule that is connected with the matched inconsistency pattern to the data post to evaluate whether there is an inconsistency in the data post or not, wherein the number of inconsistency patterns each defines an inconsistency that has previously occurred for the value of the data attribute in the data repository,
wherein each of the number of inconsistency patterns has a probability weight, the probability weight indicating a frequency of inconsistency of each of the inconsistency patterns that occurred for the data posts in the data repository; and
if there is more than one match, selecting inconsistency pattern in an inconsistency pattern order according to the probability weight for each matched inconsistency pattern.

15. The non-transitory computer-readable medium of claim 14 , wherein if there is more than one match, the inconsistency pattern with highest probability weight is selected first.

Assignees

• Gullin Patrick
• ERICSSON TELEFON AB L M (publ)

Inventors

• Gullin Patrick

Classifications

• G06N7/01

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

• G06F16/273

  Asynchronous replication or reconciliation · CPC title

• G06F16/215

  Improving data quality; Data cleansing, e.g. de-duplication, removing invalid entries or correcting typographical errors · CPC title

• G06F16/2365 Primary

  Ensuring data consistency and integrity · CPC title

• G06N5/027

  Frames · CPC title