Open j-kreis opened 3 years ago
Hi @ticarki, thanks for reporting!
It's hard to pinpoint the exact problem without the actual data. Does the following code lead to an identical error on your M1 chip?
dt <- data.frame(
A = 1:1000,
B = 1.1:10000.1
)
path <- tempfile(".fst")
fst::write_fst(dt, path)
head(fst::fst(path)[1:2000, 1:2])
#> A B
#> 1 1 1.1
#> 2 2 2.1
#> 3 3 3.1
#> 4 4 4.1
#> 5 5 5.1
#> 6 6 6.1
also, what endianness is the M1 chip?
# on intel/AMD
.Platform$endian
#> [1] "little"
thanks
Hi @MarcusKlik, thanks for your answer! Blame on me, I probably used an incomplete fst file. Would it be possible to add a check for that? Not relevant anymore, but .Platform$endian
returns "little"
Hi @ticarki, thanks for reporting back. Yes, adding a check for incomplete files would definitely add to the stability of fst
. I will change the title of your issue and change it to a feature request!
I am running R on a Mac with a M1 chip.
When I use the following code chunk R crashes with a 'floating point exception':
Reading smaller chunks of the fst works fine.