Skip to content
Back to formatted view

Raw Message

Message-ID: <Pine.LNX.4.64.0709200709460.14389@homer22.u.washington.edu>
Date: 2007-09-20T14:16:36Z
From: Thomas Lumley
Subject: Package Survey
In-Reply-To: <2F01C2D1A2A44B41BD01DDE5F72E1CAB01A1F9EF@rngpew02.drn.ine.pt>

On Thu, 20 Sep 2007, Rita Sousa wrote:

> How I use the function as.svrepdesign without memory.size problems?
>
> desenho_npc_JK <- as.svrepdesign(desenho_npc,type="JKn")
>
> Error: cannot allocate vector of size 161.3 Mb

There is currently no easy way to affect the amount of memory that this 
uses, unless you can divide the data into pieces -- for example, use only 
some variables, or create separate design objects for strata.

The surveyNG package, which can handle very large designs, does not yet do 
replicate-weight analyses.


One question is how big the design actually is.
161.3Mb looks like a vector of length 21 million

If this is the number of observations in the study then you need a 
larger computer.  If it is the total number of replicate weights then you 
may just be able to increase the memory limit for R  on your computer 
using the memory.size() function, as the warning method suggests.

 	-thomas

> In addition: Warning messages:
>
> 1: Reached total allocation of 1022Mb: see help(memory.size)
>
> 2: Reached total allocation of 1022Mb: see help(memory.size)
>
> 3: Reached total allocation of 1022Mb: see help(memory.size)
>
> 4: Reached total allocation of 1022Mb: see help(memory.size)
>
>
>
> Many thanks,
>
> ---------------------------------------------------
> Rita Sousa
> Departamento de Metodologia e Sistemas de Informa??o
>
> INE - DP: Instituto Nacional de Estat?stica - Delega??o do Porto
>
> Tel.: 22 6072016 (Extens?o: 4116)
> ---------------------------------------------------
>
>
>
>
> 	[[alternative HTML version deleted]]
>
>

Thomas Lumley			Assoc. Professor, Biostatistics
tlumley at u.washington.edu	University of Washington, Seattle