I am running uniroot function in the loop, but hit an error and the code stopped. The code is as below;
func <-function(f) -a*b/c*0.5*d*e^2 + (d/f-1)*g*sin(h*(pi/180))-i
dat <- data.frame(a = c(0.99,0.99,0.99),
b = c(0.1986572,0.1986572,0.1986572),
c = c(237.5,237.5,237.5),
d = c(1028.372, 1028.711, 1028.372),
e = c(2.46261, 2.986461, 2.46261),
f = c(-1,-1,-1),
g = c(9.8,9.8,9.8),
h = c(-54.97964, -51.65978, -54.97964),
i = c(0.03699588, -0.0375189, 0.03699588))
for(j in 1:length(dat$a)){
a <- dat$a[j]
b <- dat$b[j]
c <- dat$c[j]
d <- dat$d[j]
e <- dat$e[j]
#f: this should be solved by uniroot
g <- dat$g[j]
h <- dat$h[j]
i <- dat$i[j]
sol <- uniroot(func,c(0, 2000),extendInt = "yes")
dat$f[j] <- sol$root
print(j)
}
Running above code, hit the below error:
[1] 1
Error in uniroot(func, c(0, 2000), extendInt = "yes") :
no sign change found in 1000 iterations
The code stopped at j=1, and did not go to j=2 & 3. Therefore, dat$f shows
> dat$f
[1] 1526.566 -1.000 -1.000
My goal is when uniroot hits an error in a given j, put NA in dat$f[j], and continue the loop by the end.
If this works, dat$f[1] and dat$f[3] should have the same value (=1526.566) using the the above dataframe.
Please advice me on how to deal with the uniroot error.
(This post was modified to include the code for being reproducible for everyone, as suggested in my last post)
Try to extend the range of the interval. For instance:
sol <- uniroot(func, c(0, 5000), extendInt = "yes")
Your code stops because within the range you defined uniroot did not find a solution