R中的向量化IF语句？

Question

x <- seq(0.1,10,0.1)
y <- if (x < 5) 1 else 2

I would want the if to operate on every single case instead of operating on the whole vector. What do I have to change?

Answer 1

x <- seq(0.1,10,0.1)

> x
  [1]  0.1  0.2  0.3  0.4  0.5  0.6  0.7  0.8  0.9  1.0  1.1  1.2  1.3  1.4  1.5
 [16]  1.6  1.7  1.8  1.9  2.0  2.1  2.2  2.3  2.4  2.5  2.6  2.7  2.8  2.9  3.0
 [31]  3.1  3.2  3.3  3.4  3.5  3.6  3.7  3.8  3.9  4.0  4.1  4.2  4.3  4.4  4.5
 [46]  4.6  4.7  4.8  4.9  5.0  5.1  5.2  5.3  5.4  5.5  5.6  5.7  5.8  5.9  6.0
 [61]  6.1  6.2  6.3  6.4  6.5  6.6  6.7  6.8  6.9  7.0  7.1  7.2  7.3  7.4  7.5
 [76]  7.6  7.7  7.8  7.9  8.0  8.1  8.2  8.3  8.4  8.5  8.6  8.7  8.8  8.9  9.0
 [91]  9.1  9.2  9.3  9.4  9.5  9.6  9.7  9.8  9.9 10.0

> ifelse(x < 5, 1, 2)
  [1] 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
 [38] 1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
 [75] 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2

Answer 2

y <- if (x < 5) 1 else 2 does not operate on the whole vector (the warning you receive tells you only the first element of the condition will be used). You want ifelse :

y <- ifelse(x < 5, 1, 2)

ifelse operates on the whole logical vector, element-by-element. if only accepts one logical value. See ?"if" and ?ifelse

Answer 3

For completeness: In big vectors, you can use the indices to speed things up (we do that often in simulations, where functions typically run 1000 to 10000 times). But as long as it isn't necessary, just use ifelse . This reads a lot easier.

> set.seed(100)
> x <- runif(1000,1,10)

> system.time(replicate(10000,{
+     y <- ifelse(x < 5,1,2)
+ }))
   user  system elapsed 
   2.56    0.08    2.64 

> system.time(replicate(10000,{
+   y <- rep(2,length(x))
+   y[x < 5]<- 1
+ }))
   user  system elapsed 
   0.48    0.00    0.48 

Answer 4

You could also just create a logical vector and 1 to it

x <- seq(0.1, 10, 0.1) # Your data set   
(x >= 5) + 1
#  [1] 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
# [92] 2 2 2 2 2 2 2 2 2

---

If would like to compare performance, it would be the fastest solution

set.seed(100)
x <- runif(1e6, 1, 10)

RL <- function(x) y <- ifelse(x < 5,1,2)
JM <- function(x) {y <- rep(2, length(x)); y[x < 5] <- 1}
DA <- function(x) y <- (x >= 5) + 1

library(microbenchmark)
microbenchmark(RL(x),
               JM(x),
               DA(x))

# Unit: milliseconds
#  expr       min        lq      mean    median        uq       max neval
# RL(x) 331.83448 366.52940 378.89182 374.99741 381.08659 609.21218   100
# JM(x)  38.72894  42.18745  44.36493  43.25086  44.09626  82.76168   100
# DA(x)  10.01644  11.96482  14.21593  13.17825  14.12930  53.76923   100

Answer 5

Following the above post you can even use and modify the elements of a vector satisfying the criteria. In my opinion if it's not more costly to compute faster one should always do it.

x = seq(0.1,10,0.1)
y <- rep(2,length(x))
y[x<5] <- x[x<5]*2

The code of the previous post is best to answer the question. But if I had to use the code above I would do:

x = seq(0.1,10,0.1)
y <- rep(2,length(x))
y[x<5] <- x[x<5]*0 +1

Answer 6

nzMean <- function(x) { mean(x[x!=-1],na.rm=TRUE)}

nzMin <- function(x) {min(x[x!=-1],na.rm=TRUE)}

nzMax <- function(x) { max(x[x!=-1],na.rm=TRUE)}

nzRange<-function(x) {nzMax(x)-nzMin(x)}

nzSD <- function(x) { SD(x[x!=-1],na.rm=TRUE)}

#following function works
nzN1<- function(x) {ifelse(x!=-1,(x-nzMin(x))/nzRange(x) ,x) }

#following is bad as it returns only 4 not 5 elements of vector
nzN2<- function(x) {ifelse(x!=-1,(x[x!=-1]-nzMin(x))/nzRange(x) ,x) }

#following is bad as it returns 5 elements of vector but not correct answer
nzN3<- function(x) {ifelse(x!=-1,(x[x!=-1]-nzMin(x))/nzRange(x) ,-1) }

y<-c(1,-1,-20,2,4)
a<-nzMean(y)
b<-nzMin(y)
c<-nzMax(y)
d<-nzRange(y)
# test the working function
z<-nzN1(y)

print(z)