R help - Jan 2005 - subsetting like in SAS

Hi,

Being in the process of translating some of my SAS programs to R, I 
encountered one difficulty. I have a solution, but it is not elegant 
(and not pleasant to implement).

I have a large dataset with many variables needed to identify the 
origin of a sample, many to describe sample characteristics, others to 
describe site characteristics.

I want only a (shorter) list of sites and their characteristics.

If "origin", "ship_cat", "ship_nb",
"trip" and "set" are needed to
identify a site, in SAS you'd sort on those variables, then read the 
data with:

data sites;
	set alldata;
	by origin ship_cat ship_nb trip set;
	if first.set;
	keep list-of-variables-detailing-sites;
run;

In R I did this with the Lag function of Hmisc, and the original data 
set also needs to be sorted first:

oL <- Lag(origin)
scL <- Lag(ship_cat)
snL <- Lag(ship_nb)
tL <- Lag(trip)
sL <- Lag(set)
same <- origin==oL & ship_cat==scL & ship_nb==snL & trip==tL
& set==sL
sites <- subset(alldata, !same, 
select=c(list-of-variables-detailing-sites)

Could I do better than this?

Thanks in advance,

Denis Chabot

Hi Denis

maybe unique() can choose unique entries from your data set 
without need for sorting.

Cheers
Petr

On 13 Jan 2005 at 11:52, Denis Chabot wrote:
> Hi,
> 
> Being in the process of translating some of my SAS programs to R, I
> encountered one difficulty. I have a solution, but it is not elegant
> (and not pleasant to implement).
> 
> I have a large dataset with many variables needed to identify the
> origin of a sample, many to describe sample characteristics, others to
> describe site characteristics.
> 
> I want only a (shorter) list of sites and their characteristics.
> 
> If "origin", "ship_cat", "ship_nb",
"trip" and "set" are needed to
> identify a site, in SAS you'd sort on those variables, then read the
> data with:
> 
> data sites;
>  set alldata;
>  by origin ship_cat ship_nb trip set;
>  if first.set;
>  keep list-of-variables-detailing-sites;
> run;
> 
> In R I did this with the Lag function of Hmisc, and the original data
> set also needs to be sorted first:
> 
> oL <- Lag(origin)
> scL <- Lag(ship_cat)
> snL <- Lag(ship_nb)
> tL <- Lag(trip)
> sL <- Lag(set)
> same <- origin==oL & ship_cat==scL & ship_nb==snL & trip==tL
& set==sL
> sites <- subset(alldata, !same,
> select=c(list-of-variables-detailing-sites)
> 
> Could I do better than this?
> 
> Thanks in advance,
> 
> Denis Chabot
> 
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide!
> http://www.R-project.org/posting-guide.html
Petr Pikal
petr.pikal at precheza.cz

I want to thank Petr Pikal, Robert Balshaw and Na Li for suggesting the 
use of "unique" or "!duplicated" on a subset of my data
where unwanted
variables have been removed. This worked perfectly.

Denis Chabot
On 13 Jan 2005 at 11:52, Denis Chabot wrote:
> Hi,
>
> Being in the process of translating some of my SAS programs to R, I
> encountered one difficulty. I have a solution, but it is not elegant
> (and not pleasant to implement).
>
> I have a large dataset with many variables needed to identify the
> origin of a sample, many to describe sample characteristics, others to
> describe site characteristics.
>
> I want only a (shorter) list of sites and their characteristics.
>
> If "origin", "ship_cat", "ship_nb",
"trip" and "set" are needed to
> identify a site, in SAS you'd sort on those variables, then read the
> data with:
>
> data sites;
>  set alldata;
>  by origin ship_cat ship_nb trip set;
>  if first.set;
>  keep list-of-variables-detailing-sites;
> run;
>
> In R I did this with the Lag function of Hmisc, and the original data
> set also needs to be sorted first:
>
> oL <- Lag(origin)
> scL <- Lag(ship_cat)
> snL <- Lag(ship_nb)
> tL <- Lag(trip)
> sL <- Lag(set)
> same <- origin==oL & ship_cat==scL & ship_nb==snL & trip==tL
& set==sL
> sites <- subset(alldata, !same,
> select=c(list-of-variables-detailing-sites)
>
> Could I do better than this?