Bryan Keller
2009-Sep-15 23:51 UTC
[R] How to quickly find the position of the first non-zero element of a vector
Anyone familiar with a quicker method to find the position of the first non-zero
element of a vector?
#The first way:
print(min(which(vector != 0)))
#The second way:
for(i in 1:length(vector)) {
if (vector[i] != 0) {
print(i)
break
}
}
The first way seems to be faster for larger vectors (or when the first non-zero
is deeper in)...
vlarge <- c(numeric(20),1:20)
#Method 1:
system.time(
for (i in 1:30000) {
min(which(vlarge != 0))})
# user system elapsed
# 0.49 0.00 0.48
#Method 2:
system.time(
for (i in 1:30000) {
for (i in 1:40) {
if (vlarge[i] != 0) {
break
}
}})
# user system elapsed
# 0.99 0.02 0.99
--------------------------------------------------
The second way seems to be faster for smaller vectors (or when the first
non-zero is closer to the front)...
vsmall <- c(numeric(5),1:5)
#Method 1:
system.time(
for (i in 1:30000) {
min(which(vsmall != 0))})
# user system elapsed
# 0.41 0.00 0.42
#Method 2:
system.time(
for (i in 1:30000) {
for (i in 1:10) {
if (vsmall[i] != 0) {
break
}
}})
# user system elapsed
# 0.31 0.02 0.31
Or, might the fastest way be to choose between the two methods on-the-fly based
on length of the vector, etc.?
Bryan
-------------
Bryan Keller, Doctoral Student/Project Assistant
Educational Psychology - Quantitative Methods
The University of Wisconsin - Madison
jim holtman
2009-Sep-16 00:55 UTC
[R] How to quickly find the position of the first non-zero element of a vector
try .which.mine Here is what is does on my system:> vlarge <- c(numeric(20),1:20) > system.time(+ for (i in 1:30000) { + which.min(vlarge != 0)}) user system elapsed 0.30 0.00 0.35> > #Method 2: > system.time(+ for (i in 1:30000) { + for (i in 1:40) { + if (vlarge[i] != 0) { + break + } + }}) user system elapsed 2.25 0.01 2.65>On Tue, Sep 15, 2009 at 7:51 PM, Bryan Keller <bskeller at wisc.edu> wrote:> Anyone familiar with a quicker method to find the position of the first non-zero element of a vector? > > #The first way: > print(min(which(vector != 0))) > > #The second way: > for(i in 1:length(vector)) { > ? ? ? ?if (vector[i] != 0) { > ? ? ? ? ? ? ? ?print(i) > ? ? ? ? ? ? ? ?break > ? ? ? ? ? ? ? ?} > ? ? ? ?} > > The first way seems to be faster for larger vectors (or when the first non-zero is deeper in)... > > vlarge <- c(numeric(20),1:20) > > #Method 1: > system.time( > for (i in 1:30000) { > ? ? ? ?min(which(vlarge != 0))}) > # ?user ?system elapsed > # ?0.49 ? ?0.00 ? ?0.48 > > #Method 2: > system.time( > for (i in 1:30000) { > for (i in 1:40) { > ?if (vlarge[i] != 0) { > ? ?break > ?} > }}) > # ?user ?system elapsed > # ?0.99 ? ?0.02 ? ?0.99 > > -------------------------------------------------- > The second way seems to be faster for smaller vectors (or when the first non-zero is closer to the front)... > > vsmall <- c(numeric(5),1:5) > > #Method 1: > system.time( > for (i in 1:30000) { > ? ? ? ?min(which(vsmall != 0))}) > # ?user ?system elapsed > # ?0.41 ? ?0.00 ? ?0.42 > > #Method 2: > system.time( > for (i in 1:30000) { > for (i in 1:10) { > ?if (vsmall[i] != 0) { > ? ?break > ?} > }}) > # ?user ?system elapsed > # ?0.31 ? ?0.02 ? ?0.31 > > Or, might the fastest way be to choose between the two methods on-the-fly based on length of the vector, etc.? > > Bryan > > > ------------- > Bryan Keller, Doctoral Student/Project Assistant > Educational Psychology - Quantitative Methods > The University of Wisconsin - Madison > > ______________________________________________ > R-help at r-project.org mailing list > https://stat.ethz.ch/mailman/listinfo/r-help > PLEASE do read the posting guide http://www.R-project.org/posting-guide.html > and provide commented, minimal, self-contained, reproducible code. >-- Jim Holtman Cincinnati, OH +1 513 646 9390 What is the problem that you are trying to solve?
jim holtman
2009-Sep-16 00:58 UTC
[R] How to quickly find the position of the first non-zero element of a vector
Fast fingers. Had to change the test:> vlarge <- c(numeric(20),1:20) > system.time(+ for (i in 1:30000) { + which.min(!(vlarge != 0))}) # test changed user system elapsed 0.35 0.00 0.42> > #Method 2: > system.time(+ for (i in 1:30000) { + for (i in 1:40) { + if (vlarge[i] != 0) { + break + } + }}) user system elapsed 2.43 0.00 2.78>On Tue, Sep 15, 2009 at 7:51 PM, Bryan Keller <bskeller at wisc.edu> wrote:> Anyone familiar with a quicker method to find the position of the first non-zero element of a vector? > > #The first way: > print(min(which(vector != 0))) > > #The second way: > for(i in 1:length(vector)) { > ? ? ? ?if (vector[i] != 0) { > ? ? ? ? ? ? ? ?print(i) > ? ? ? ? ? ? ? ?break > ? ? ? ? ? ? ? ?} > ? ? ? ?} > > The first way seems to be faster for larger vectors (or when the first non-zero is deeper in)... > > vlarge <- c(numeric(20),1:20) > > #Method 1: > system.time( > for (i in 1:30000) { > ? ? ? ?min(which(vlarge != 0))}) > # ?user ?system elapsed > # ?0.49 ? ?0.00 ? ?0.48 > > #Method 2: > system.time( > for (i in 1:30000) { > for (i in 1:40) { > ?if (vlarge[i] != 0) { > ? ?break > ?} > }}) > # ?user ?system elapsed > # ?0.99 ? ?0.02 ? ?0.99 > > -------------------------------------------------- > The second way seems to be faster for smaller vectors (or when the first non-zero is closer to the front)... > > vsmall <- c(numeric(5),1:5) > > #Method 1: > system.time( > for (i in 1:30000) { > ? ? ? ?min(which(vsmall != 0))}) > # ?user ?system elapsed > # ?0.41 ? ?0.00 ? ?0.42 > > #Method 2: > system.time( > for (i in 1:30000) { > for (i in 1:10) { > ?if (vsmall[i] != 0) { > ? ?break > ?} > }}) > # ?user ?system elapsed > # ?0.31 ? ?0.02 ? ?0.31 > > Or, might the fastest way be to choose between the two methods on-the-fly based on length of the vector, etc.? > > Bryan > > > ------------- > Bryan Keller, Doctoral Student/Project Assistant > Educational Psychology - Quantitative Methods > The University of Wisconsin - Madison > > ______________________________________________ > R-help at r-project.org mailing list > https://stat.ethz.ch/mailman/listinfo/r-help > PLEASE do read the posting guide http://www.R-project.org/posting-guide.html > and provide commented, minimal, self-contained, reproducible code. >-- Jim Holtman Cincinnati, OH +1 513 646 9390 What is the problem that you are trying to solve?