Ruby and RDBS

[Thilanka Marasinghe]

[Note: parts of this message were removed to make it a legal post.]

This may seem like a very newb question. BUt are there any groups or
projects dealing with using ruby and RDBMS, to do data mining and data
analysis?

I'm trying to learn ruby and want to am into database and data mining
development and am curious if I could combine both together

cheers

Thilankka

 

[David Masover]

I'm trying to learn ruby and want to am into database and data mining
development and am curious if I could combine both together

The Rails community seems to want to avoid writing any SQL at all, and would
rather work with document databases (things like ThruDB, SimpleDB, and
BigTable).

That said... Yes, absolutely, depending on the database. Ruby has adapters for
most major database engines. You're probably going to want to learn some
SQL...

A quick Google for "ruby database mining" shows something called Ruminate, but
I can't find any active development lately -- so if you're looking for a
prebuilt solution, you're probably on your own.

 

[Thilanka Marasinghe]

[Note: parts of this message were removed to make it a legal post.]

oh thanks.. Maybe I should hav a look at these document databases as well

 

[Farrel Lifson]

This may seem like a very newb question. BUt are there any groups or
projects dealing with using ruby and RDBMS, to do data mining and data
analysis?

Sequel (http://sequel.rubyforge.org/) would suit you well. It's a nice
Ruby like interface to a number of different databases and can act as
a plan Ruby -> SQL interface or use an Object Relational Mapping.

Farrel

 

[Anthony Eden]

The Rails community seems to want to avoid writing any SQL at all, and would
rather work with document databases (things like ThruDB, SimpleDB, and
BigTable).

Eh? ActiveRecord is tightly tied to SQL. The fact that it provides an
interface that attempts to minimize the use of SQL doesn't change the
fact that underneath it all is heaps and heaps of SQL that is
generated. Additionally BigTable and SimpleDB are not document
databases.

Sincerely,
Anthony Eden

 

[Anthony Eden]

This may seem like a very newb question. BUt are there any groups or
projects dealing with using ruby and RDBMS, to do data mining and data
analysis?

Data mining and data analysis are pretty broad and can be accomplished
with a variety of techniques. Perhaps you can narrow it down a little
bit?

For example, if you are looking to use dimensional modeling for data
warehousing and analysis then you might want to take a look at
ActiveWarehouse. For data mining you might want to consider search
with something like Solr or Sphinx and distributed techniques such as
map/reduce. What you will use depends largely on the amount of data
you are trying to analyze and the resources available to you.

Sincerely,
Anthony Eden

 

[James Gray]

The Rails community seems to want to avoid writing any SQL at all,
and would
rather work with document databases (things like ThruDB, SimpleDB, and
BigTable).

I hear comments like this a lot, but I don't really understand where
they are coming from.

ActiveRecord, the ORM layer of Rails, is capable of dropping down to
whatever level of SQL you need. Many of the features added to it over
time have been to use SQL more heavily for various advantages. Take
"Eager Loading," for example.

I program in Rails all day every day and if you removed all the SQL I
use, I would be much worse at it.

James Edward Gray II

 

[David Masover]

Eh? ActiveRecord is tightly tied to SQL.

Yes... Maybe I was being too bold.

However, ActiveRecord seems to be written from the perspective of hiding SQL
from the user. It's there if you need it, but for a lot of the concepts, it's
an implementation detail.

The structure of STI and polymorphism in Rails suggests that there might at
least be SQL-like databases targeting ActiveRecord in the future.

Additionally BigTable and SimpleDB are not document
databases.

They do seem to follow the same idea, though, and BigTable is certainly used
like one.

My point was that at least some parts of the community seem to embrace the
idea of a document database as a way to scale. And yes, it requires replacing
ActiveRecord -- but ActionPack should still be useful.