Message-ID: <20030518222741.GA3923@sonny.eddelbuettel.com>
Date: 2003-05-18T22:27:41Z
From: Dirk Eddelbuettel
Subject: How to split a dataframe into smaller constituent dataframes
In-Reply-To: <Law12-F62gRzpPWywCP00029bf5@hotmail.com>
On Sun, May 18, 2003 at 10:11:11PM +0000, Neil Osborne wrote:
> I have read a large dataset into a dataframe using RODBC, the rows of data
> in the dataframe are (integer) timestamped and I would like to divide the
> original dataframe into n smaller dataframes where dataframe 1 contains all
> rows that had timestamps falling in the period 0-x1 minutes, dataframe 2
> contains all rows that had timestamps falling between x1+1 and x2, etc..
> Does anyone know how to do this?
I would convert all your dates to DateTimeClass (i.e. POSIXct or POSIXlt),
e.g. using strptime, and then loop over them doing your subsetting at each
grid step you specify. R really has excellent facilities to compute on, or
compare, dates up to Unix timestamp precision.
Alternatively, if the data lives in a DBMS you could simply tailor your
queries to the desired grid sizes and retrieve individual chunks and store
those.
Hth, Dirk
--
Don't drink and derive. Alcohol and analysis don't mix.