In the introduction of Computational Actuarial Science with R, there was a short paragraph on how could we import only some parts of a large database, by selecting specific variables. The trick was to use the following

> read.table.select.columns=function(datatablename,
I,sep=";"){
+ datanc=read.table(datatablename,header=TRUE,
sep=sep,skip=0,nrows=1)
+ mycols=rep("NULL",ncol(datanc))
+ names(mycols)=names(datanc)
+ mycols[I]=NA
+ datat=read.table(datatablename,header=TRUE,
sep=sep,colClasses=mycols)
+ return(datat)}

For instance, if we use the same dataset as in the introduction, we can import only two variables of interest,

> loc="http://myweb.fsu.edu/jelsner/extspace/extremedatasince1899.csv"
> dt1=read.table.select.columns(loc,c("Region",
"Wmax"),sep=",")
> head(dt1,10)
    Region      Wmax
1    Basin 105.56342
2    Basin  40.00000
3    Basin  35.41822
4    Basin  51.06743
5  Florida  87.34328
6    Basin  96.64138
7     Gulf  35.41822
8       US  35.41822
9       US  87.34328
10      US 106.35318
> dim(dt1)
[1] 2100    2

Continue reading How to import some parts of a large database →