| 摘要 |
A method of determining the nature of the contents of a data field. A field identification module (FIM) 102, scans the contents of the database 106 to determine a data field identifier for each of the data fields to be worked upon. The FIM 102 then compares each of the data field identifiers with a list of known identifiers stored in a repository 108 and assess a first score in respect of the comparison wherein the score is representative of a probability that the comparison is an accurate comparison. When a first probability score has been determined, the FIM 102 samples the data field according to a preselected list of requirements and make a second comparison with the data field. The FIM 102 then assesses a second score. After sampling the FIM 102 constructs a test case wherein the data field is used in the host application and a third comparison is made in respect of the test case. The FIM 102 then assesess a third score. The FIM 102 chooses a field type based on the weighted average result of the three scores and confirmed by a system operator. Once confirmed, the data field identifier and the data field represented thereby are added to the repository 108 as one of the known identifiers. The host application may be a postal coding system.
|
