|
|
This concludes the description of using data rules to find evidence of inaccurate data. This step is not easy and requires staff that is very good at formulating correct and efficient queries. And yet, the content is mostly business oriented. It again calls for a team effort between good technical staff and good business staff.
At this point in data profiling you have completed all of the testing that defines explicit data rules. If you have done everything as described, you have done a lot of work and have a very solid handle on the accuracy of your metadata and the accuracy of your data. The next chapter describes how to use other types of computations to indicate the presence of inaccurate data. They are not rules in that they do not specify what the data should say. They are computations that will alert you to problems if their results are too far from being reasonable.
|
|