[R-SIG-Finance] get.hist.quotes and MySQL

Brian G. Peterson brian at braverock.com
Fri Sep 21 12:00:49 CEST 2007

paul sorenson wrote:
> I am collecting hundreds of time series and storing them in sqlite.  I 
> am downloading and storing with python then reading them out with R.
> I don't know if the MySQL and PostgreSQL interfaces are smarter, but 
> with SQLite the data all comes out as character, you need to perform 
> some massaging after extracting data from SQLite.  On the other hand 
> SQLite is zero administration.

Just for the archive:  yes MySQL, PostgreSQL, Oracle, DB2, etc can store 
data types better than sqlite.  As Paul notes, the main reason to use 
sqlite is that you don't need to actually administer a real database 

   - Brian

> Dry Eraser wrote:
>> I want to screen the universe of stock price histories available from
>> Yahoo for certain patterns.  That means iterating a screen over
>> several thousand time series.  get.hist.quotes saves time series to
>> CSV.  Putting everything into a database like MySQL seems to be a
>> better idea than working with thousands of CSV files, especially when
>> I might want to add new data points in the future.
>> I scoured the R and R packages manuals, the TOC of the Times Series
>> Analysis Using R book, and googled r lists for information on
>> get.hist.quotes and MySQL and came up only almost nothing on the
>> practical use of MySQL.  (I did find a tantalizing pointer by
>> Bollinger to a bunch of python, r and gnuplot packages for his own
>> projects.  I think I might need to leverage python or perl to load up
>> the database and then R packages manipulate the data.)
>> Any suggestions on where to learn more -- and find examples -- on
>> using MySQL for ts and any comments on my type of project using R in
>> general?
>> Thanks, Ming

More information about the R-SIG-Finance mailing list