Abstract
A method for cleaning data records in a database comprises receiving a plurality of quality constraints, parsing the data records in the database to detect data records that violate at least one of the quality constraints and determining modifications that are required to repair at least some of the data records that violate at least one of the quality constraints. The method further comprises processing the determined modifications using a satisfiability (SAT) solver to identify modifications that repair a plurality of violations simultaneously, generating a repair context comprising expressions for performing each the modifications identified by the satisfiability solver and applying the modifications in the repair context to the database to clean at least some of the data records in the database.
| Original language | English |
|---|---|
| Patent number | WO2014000785 |
| IPC | G06F 17/ 30 A I |
| Priority date | 27/06/12 |
| Publication status | Published - 3 Jan 2014 |