Friday 22 July 2011

Data cleansing

Usually the first thing people want to do when they want better access to their data is to build a database. That's all good, and I understand the desire, but the implied result is that the data will be more reliable once it's entered. This is never the case. As a result of any automatic process, the data will only decay unless a significant effort has gone into examining and repairing it first, which is called data cleansing. That's usually what people want when they propose to Make A Database, but the long hard data cleansing work is separate and quite different from that. It can be done without making a database first, but it's easier once the database structures are in place.

Mokalus of Borg

PS - It can be quite a lot of work.
PPS - Especially if there's a lot of data.

No comments: