Date: Thu, 4 Feb 1999 18:51:29 -0300
Reply-To: "Hector E. Maletta" <hmaletta@OVERNET.COM.AR>
Sender: "SPSSX(r) Discussion" <SPSSX-L@UGA.CC.UGA.EDU>
From: "Hector E. Maletta" <hmaletta@OVERNET.COM.AR>
Subject: Re: recoding date of birth
Content-Type: text/plain; charset=us-ascii
Jeffrey Quon wrote:
> Apologies in advance if this question proves too elementary, however...
> I would like to help a colleague analyse a large data set (containing
> over 4600 cases).
Yours, Jeff, is a fairly small data set, even if each case contains dozens of
variables. In fact, SPSS limitations are so high they are for all practical purposes
non-existent (there is an upper limit of 4 Gb for the size of data files, or at least
that was the limit for version 6.1; I suppose it is still so for more recent versions).
You can manage millions of cases and thousands of variables if your hard disk would
accommodate the original data, the working file and the eventually transformed data
file, i.e. about thrice the size of your data.
I am presently working with four years of the Brazilian National Household survey,
comprising more than 300,000 individuals in about 90,000 households per year (that is
about 1.2 million cases in the whole data set) with some 300 variables totalling about
800 Mb of disk space (when saved in compressed form). My 6 Gb disk does not complain.
So don't worry about the size of your data set.
Universidad del Salvador
Buenos Aires, Argentina