Re: How to efficiently make an "history" ?

From: Theo Peterbroers <peterbroers_at_floron.leidenuniv.nl>
Date: 30 Aug 2001 07:44:16 -0700
Message-ID: <39bb2c10.0108300644.5ff9cf75_at_posting.google.com>

"kesako" <kesako_at_mail.com> wrote in message news:<9ml0jg$ojd$1_at_news1.span.ch>...
> Hello Theo,

<SNIP>

> Sorry for creating confusion.
>
> > - updates of existing entities and insertion of new ones are
> > undiscernable.
>
> Not following here. I agree that we need to put business rules on how to use
> the schema correctly. Do we speak about the same thing?

Your description of the table is:
(field_name, field_value, revision_comment, revision_time) Example:

('Name', 'John', 'Hi there', 2000083000987etc.)
('Place', 'Paris', 'I mean Paris, Texas', 2000083000988etc.)
('Name', 'Adam', 'Nice name',2000083000989etc.)

Question: Did John change his name to Adam or is Adam a new person?

> > In order to decide which row belongs to which entity, you need to
> > create another column identifiyn entities:
(entityID, field_name, field_value, revision_comment, revision_time) Example:

(ID001, 'Name', 'John', 'Hi there', 2000083000987etc.)
(ID002, 'Place', 'Paris', 'I mean Paris, Texas', 2000083000988etc.)
(ID001, 'Name', 'Adam', 'Nice name',2000083000989etc.)

Aha, John did change his name.
(Id003, 'Name', 'Adam', 'Nice name',2000083000990etc.) Now there are two Adams. Received on Thu Aug 30 2001 - 16:44:16 CEST