memisc (version 0.11-9)

aggregate.formula: Convenient and flexible aggregation of data frames

Description

aggregate.formula constructs a data frame of summaries conditional on given values of independent variables given by a formula. It is a method of the generic function aggregate applied to formula objects. genTable does the same, but produces a table.

Usage

## S3 method for class 'formula':
aggregate(x, data=parent.frame(), subset=NULL, na.action, exclude = c(NA, NaN),
      drop.unused.levels = FALSE, names=NULL, addFreq=TRUE,...)
genTable(formula, data=parent.frame(), subset=NULL, na.action, exclude = c(NA, NaN),
      drop.unused.levels = FALSE, names=NULL, addFreq=TRUE)

Arguments

x, formula

a formula object with an expression yielding a numeric result on the left hand side and the conditioning variables, separated by +, on the right hand side. Interactions are ignored. The left hand side of the formula is optiona

data

an environment or data frame or an object coercable into a data frame.

subset

an optional vector specifying a subset of observations to be used.

na.action

a function which indicates what should happen when the data contain NAs.

exclude

a vector of values to be excluded when forming the set of levels of the classifying factors.

drop.unused.levels

a logical indicating whether to drop unused levels in the classifying factors.

names

an optional character vector giving names to the result(s) yielded by the expression on the left hand side of formula. This argument may be redundant if the left hand results in is a named vector. (See the example below.)

addFreq

a logical value. If TRUE and data is a table or a data frame with a variable named "Freq", a call to table, wtable, or percen

...

further arguments, ignored.

Value

• aggregate.formula results in a data frame with conditional summaries and unique value combinations of conditioning variables. genTable returns a table, that is, an array with class "table".

Details

If an expression is given as left hand side of the formula, its value is computed for any combination of values of the values on the right hand side. If the right hand side is a dot, then all variables in data are added to the right hand side of the formula.

If no expression is given as left hand side, then the frequency counts for the respective value combinations of the right hand variables are computed.

If a single factor is on the left hand side, then the left hand side is translated into an appropriate call to table(). Note that also in this case addFreq takes effect. If a single numeric variable is on the left hand side, frequency counts weighted by this variable are computed. In these cases, aggregate.formula is equivalent to as.data.frame(xtabs(...)), but far less memory intensive if the data set or the number of different combinations of the conditioning variables is very large. Another difference is that conditioning variables are not coerced into factors.

See Also

aggregate.data.frame, xtabs

Examples

ex.data <- expand.grid(mean=c(0,100),sd=c(1,10))[rep(1:4,rep(100,4)),]
ex.data <- transform(ex.data,x=rnorm(n=nrow(ex.data),mean=ex.data$mean,sd=ex.data$sd))

aggregate(~mean+sd,data=ex.data)
aggregate(mean(x)~mean+sd,data=ex.data)
aggregate(mean(x)~mean+sd,data=ex.data,name="Average")
aggregate(c(mean(x),sd(x))~mean+sd,data=ex.data)
aggregate(c(Mean=mean(x),StDev=sd(x),N=length(x))~mean+sd,data=ex.data)
genTable(c(Mean=mean(x),StDev=sd(x),N=length(x))~mean+sd,data=ex.data)
attach(ex.data)
aggregate(c(Mean=mean(x),StDev=sd(x))~mean+sd)
genTable(c(Mean=mean(x),StDev=sd(x))~mean+sd)
detach(ex.data)
aggregate(wtable(Admit,Freq)~.,data=UCBAdmissions)
aggregate(Admit~.,data=UCBAdmissions)
aggregate(percent(Admit)~.,data=UCBAdmissions)
aggregate(percent(Admit)~Gender,data=UCBAdmissions)
aggregate(percent(Admit)~Dept,data=UCBAdmissions)
aggregate(percent(Gender)~Dept,data=UCBAdmissions)
aggregate(percent(Admit)~Dept,data=UCBAdmissions,Gender=="Female")
genTable(percent(Admit)~Dept,data=UCBAdmissions,Gender=="Female")