公开(公告)号：DE10134229A1

公开(公告)日：2002-02-28

申请号：DE10134229

申请日：2001-07-13

Applicant: IBM

Inventor： ARNING ANDREAS ,  BOLLINGER TONI ,  KEULER REINHOLD ,  SCHWENKREIS FRIEDEMANN

IPC: G06F17/30

Abstract: Determination of differences in a database table using a method based on use of a classification tree after selection of a column as a classification column. The invention also relates to a data processing program and a computer program product for use in finding differing data in a database table. Method has the following steps: selection of a column as a classification column, execution of a classification method by calculation of a classification tree with reference to the reference column. Each edge of the classification tree is assigned a grade and the leaf nodes are assigned a leaf data set that includes a part of the data records for which an evaluation of the class value equal to TRUE is determined along the whole path from the root node of the classification tree to the leaf node in question. The leaf nodes are assigned a leaf characteristic that represents the expected value in the classification column. In a final step data records differing from the leaf characteristic are classed as a difference data set.