[Tutor] use case for graphs and generators?

[Serdar Tumgoren]

Hi everyone,
I was hoping someone could advise on whether I'm tackling a specific
problem in the correct manner.

Specifically, I'm trying to use a "clean" set of historical data to
fix omissions/errors among a stream of newer data. To do so, I've
devised a series of backend SQL statements which need to be triggered
in various orders depending on the state of my object's attributes.

If my Person object, for example, only has a name and a state, then I
need to execute a SQL statement relying on those attributes to pull
additional information about the Persion. And if that search yields a
unique id for my Person object, then I can fire another SQL statement
that uses the unique ID to pull more attributes from my historical
data. Or, if my Person object happened to be instantiated with a
unique ID, then I can just skip the name/state search (and various
others) and go straight to the ID search.

After each query, I clearly need to inspect the status of the
attributes: once they've all been corrected, I can stop executing
corrective queries and move on to the next Person object.

To implement the above, I was starting to research generators and
graph data structures (e.g.
http://www.python.org/doc/essays/graphs.html).

So my question -- is that indeed the correct approach for what I've
described as the use case? I've never tackled graphs or generators
before, and I figure I have plenty of reading to do before I could
devise a solution relying on both. I'd be grateful if anyone could
weigh in on my approach, or perhaps advise on alternatives.

Regards,
Serdar