I want to convert current time to epoch timestamp with microseconds in R. For that I am running following code.
options(digits.secs = 6)
now = Sys.time()
as.double(now)
When I simply print now the output is "2017-11-11 20:40:18.370305 CET" where it is capturing microseconds but when I convert it to epoch timestamp using as.double(now), the microsecond precision is gone. The result is only 1510429218. How to retain the microseconds?
Learned how R handles decimal digits very hard way! See following code.
options(digits.secs = 6)
a = c(as.double(Sys.time()), as.double(Sys.time()), as.double(Sys.time()))
print(a)
The output of this code is following.
[1] 1510430216 1510430216 1510430216
Timestamps seem to have lost their microsecond precision. But when the data is written to file or exported it the decimal digits are visible.
df = data.frame(a)
write.csv(x = df, file = "test.csv")
Contents of the file are:
1 1510430215.888998
2 1510430215.889942
3 1510430215.889942
To also see the decimal digits in R session/console, visibility options need to be changed.
options(digits = 16)
print(a)
Now all the decimal digits are visible in R session too!
[1] 1510430215.888998 1510430215.889942 1510430215.889942
The default display digits are only 7. So if the number is greater than that, R simply chooses to discard decimal digits from displaying! It is ridiculous as it creates massive confusion. At least one or two decimal digits should be kept.
You do (always) have them, but you need to tell print to use more digits:
R> n <- Sys.time()
R> print(as.numeric(n), digits=16) ## digits matters here
[1] 1510432192.721936
R> n
[1] "2017-11-11 14:29:52.721935 CST"
R>
