fifer (version 1.1)

last.sample: Return only one row per ID

Description

Often times, an individual has multiple observations in the dataset. last.sample will loop through the entire dataset and return only one observation per individual, giving the first (or last) draw for a person, or performing some function on the variable of interest.

Usage

last.sample(ID, sort.var = NULL, decreasing = TRUE, data, FUN = NULL,
  fun.var = NULL, ...)

Arguments

ID

The name of the unique identifier, expressed as a string (e.g., "Match.Group")

sort.var

The variable to be sorted on in order to take the first (or last) sample, expressed as a string.

decreasing

How should the sort.var variable be sorted? Defaults to T.

data

The dataset with multiple rows per individual

FUN

What should be done with the multiple samples? This function can be used to extract the last (or first) sample using the decreasing/sort.var options, or a function can be performed (such as the mean) on one or more columns. See examples.

fun.var

If FUN if not null, the variable (or a vector of variables), expressed as strings to have the function applied to.

...

Other arguments passed to the chosen function.

Value

a new dataframe containing one row per ID

Examples

#### take only group 2 values
last.sample(ID="ID", sort.var="group", data=sleep)
#### take only group 1 values
last.sample(ID="ID", sort.var="group", decreasing=FALSE,data=sleep)
#### average group 1 and 2 values
last.sample(ID="ID", data=sleep, FUN=mean, fun.var="extra")
#### take the maximum extra value
last.sample(ID="ID", data=sleep, FUN=max, fun.var="extra")
#### take the mean of two columns extra value
sleep$group = as.numeric(as.character(sleep$group))
last.sample(ID="ID", data=sleep, FUN=mean, fun.var=c("group","extra"))