The BF turns from Inf to NaN with increasing Test value in a one-sample t-test

[koenderks]

Sometimes, depending on the test value in a one-sample t-test, the Bayes factor turns from Inf to NaN. An example can be found below. I'm not sure how to handle this on the R side though, maybe you have some ideas?

data("sesamesim")

set.seed(100)
testValue <- 36
test <- bain::t_test(x = sesamesim$age)
b1 <- bain::bain(x = test, hypothesis = paste0("x=", testValue, "; x>", testValue, "; x<", testValue), fraction = 1)
BF_02_1 <- b1$BFmatrix[1,3]
BF_02_1

[1] 718.2452

set.seed(100)
testValue <- 35.5
test <- bain::t_test(x = sesamesim$age)
b2 <- bain::bain(x = test, hypothesis = paste0("x=", testValue, "; x>", testValue, "; x<", testValue), fraction = 1)
BF_02_2 <- b2$BFmatrix[1,3]
BF_02_2

[1] Inf

set.seed(100)
testValue <- 35
test <- bain::t_test(x = sesamesim$age)
b3 <- bain::bain(x = test, hypothesis = paste0("x=", testValue, "; x>", testValue, "; x<", testValue), fraction = 1)
BF_02_3 <- b3$BFmatrix[1,3]
BF_02_3

[1] NaN

See also https://github.com/jasp-stats/jasp-test-release/issues/1252