1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
|
\name{cases}
\alias{cases}
\title{Distinguish between Cases Specified by Logical Conditions}
\description{
\code{cases} allows to distinguish several cases defined logical
conditions. It can be used to code these cases into a vector. The
function can be considered as a multi-condition generalization of
\code{\link{ifelse}}.
}
\usage{
cases(\dots,check.xor=c("warn","stop","ignore"),
.default=NA,.complete=FALSE,
check.na=c("warn","stop","ignore"),
na.rm=TRUE)
}
\arguments{
\item{\dots}{A sequence of logical expressions or assignment expressions containing
logical expressions as "right hand side".}
\item{check.xor}{character (either \code{"warn"}, \code{"stop"}, or \code{"ignore"})
or logical; if \code{TRUE} or equal to \code{"stop"} or
\code{"warn"},
\code{cases} checks whether the case
conditions are mutually exclusive. If this is not
satisfied and \code{check.xor} equals \code{"warn"} (the default), a warning is shown,
otherwise an error exception is raised.}
\item{.default}{a value to be used for unsatisfied conditions.}
\item{.complete}{logical, if \code{TRUE} an additional factor level is
created for the unsatisfied conditions.}
\item{check.na}{character (either \code{"warn"}, \code{"stop"}, or \code{"ignore"})
or logical; if \code{TRUE} or equal to \code{"stop"} or
\code{"warn"}, \code{cases} checks, whether any of the case
conditions evaluates to \code{NA}.
If that case, if \code{check.na} is \code{TRUE} or equals
\code{"stop"} an error exception is raised, while if \code{check.na}
equals \code{"warn"} (the default) a warning is shown.}
\item{na.rm}{a logical value; how to handle \code{NA}s (if they do not
already lead to an error exception). If \code{FALSE} if \emph{any} of the
conditions evaluates to \code{NA}, the corresponding value of the
result vector is \code{NA}. If \code{TRUE} (the default), the
resulting vector or factor is \code{NA} only for instances where all
conditions result in \code{NA}.
}
}
\details{
There are two distinct ways to use this function. Either the
function can be used to construct a factor that represents
several logical cases or it can be used to conditionally
evaluate an expression in a manner similar to \code{\link{ifelse}}.
For the first use, the \code{\dots} arguments have to be a series of
logical expressions. \code{cases} then returns a factor
with as many levels as logical expressions given as
\code{\dots} arguments. The resulting factor will attain its
first level if the first condition is TRUE, otherwise it will attain its
second level if the second condition is TRUE, etc.
The levels will be named after the conditions or, if name tags are
attached to the logical expressions, after the tags of the expressions.
Not that the logical expressions all need to evaluate to logical vectors
of the same length, otherwise an error condition is raised.
If \code{.complete} is \code{TRUE} then an additional factor level is
created for the conditions not satisfied for any of the cases.
For the second use, the \code{\dots} arguments have to be a series
of assignment expression of the type \code{<expression> <- <logical expression>}
or \code{<logical expression> -> <expression>}. For cases
in which the first logical expression is TRUE, the result of first expression that
appears on the other side of the assignment operator become elements of the
vector returned by \code{cases}, for cases in which the second logical expression is TRUE,
the result of the second expression that appears on the other side
of the assignment operator become elements of the
vector returned by \code{cases}, etc.
For cases that do not satisfy any of the given conditions the value of
the \code{.default} argument is used.
Note that the logical expressions also here all need to evaluate to logical
vectors of the same length. The expressions on the other side of the
assignment operator should also be either vectors of the same length
and mode or should scalars of the same mode, otherwise unpredictable
results may occur.
}
\value{
If it is called with logical expressions as \dots arguments,
\code{cases} returns a factor, if it is called with
assignment expressions the function returns a vector with the
same mode as the results of the "assigned" expressions
and with the same length as the logical conditions.
}
\examples{
# Examples of the first kind of usage of the function
#
df <- data.frame(x = rnorm(n=20), y = rnorm(n=20))
df <- df[do.call(order,df),]
(df <- within(df,{
x1=cases(x>0,x<=0)
y1=cases(y>0,y<=0)
z1=cases(
"Condition 1"=x<0,
"Condition 2"=y<0,# only applies if x >= 0
"Condition 3"=TRUE
)
z2=cases(x<0,(x>=0 & y <0), (x>=0 & y >=0))
}))
xtabs(~x1+y1,data=df)
dd <- with(df,
try(cases(x<0,
x>=0,
x>1,
check.xor=TRUE)# let's be fussy
)
)
dd <- with(df,
try(cases(x<0,x>=0,x>1))
)
genTable(range(x)~dd,data=df)
# An example of the second kind of usage of the function:
# A construction of a non-smooth function
#
fun <- function(x)
cases(
x==0 -> 1,
abs(x)> 1 -> abs(x),
abs(x)<=1 -> x^2
)
x <- seq(from=-2,to=2,length=101)
plot(fun(x)~x)
# Demo of the new .default and .complete arguments
x <- seq(from=-2,to=2)
cases(a = x < -1,
b = x > 1,
.complete = TRUE)
cases(x < -1,
x > 1,
.complete = TRUE)
cases(1 <- x < -1,
3 <- x > 1,
.default = 2)
threshhold <- 5
d <- c(1:10, NaN)
d1 <- cases(
d > threshhold -> 1,
d <= threshhold -> 2
)
d2 <- cases(
is.na(d) -> 0,
d > threshhold -> 1,
d <= threshhold -> 2
)
# Leads to missing values because some of the conditions result in missing
# even though they could be 'captured'
d3 <- cases(
is.na(d) -> 0,
d > threshhold -> 1,
d <= threshhold -> 2,
na.rm=FALSE
)
d4 <- cases(
is.na(d) -> 0,
d > threshhold +2 -> 1,
d <= threshhold -> 2,
na.rm=FALSE
)
cbind(d,d1,d2,d3,d4)
cases(
d > threshhold,
d <= threshhold
)
cases(
is.na(d),
d > threshhold,
d <= threshhold
)
cases(
d > threshhold,
d <= threshhold,
.complete=TRUE
)
cases(
d > threshhold + 2,
d <= threshhold,
.complete=TRUE
)
}
\keyword{manip}
|
