Nanosecond-Resolution Time for R

Full 64-bit resolution date and time support with resolution up to nanosecond granularity is provided, with easy transition to and from the standard 'POSIXct' type.


Nanosecond Time Resolution for R

R has excellent tools for dates and times. The Date and POSIXct classes (as well as the 'wide' representation in POSIXlt) are versatile, and a lot of useful tooling has been built around them.

However, POSIXct is implemented as a double with fractional seconds since the epoch. Given the 53 bits accuracy, it leaves just a bit less than microsecond resolution. Which is good enough for most things.

But more and more performance measurements, latency statistics, ... are now measured more finely, and we need nanosecond resolution. For which commonly an integer64 is used to represent nanoseconds since the epoch.

And while R does not a native type for this, the bit64 package by Jens Oehlschlägel offers a performant one implemented as a lightweight S3 class. So this package uses this integer64 class, along with two helper functions for parsing and formatting, respectively, at nano-second resolution from the RcppCCTZ package which wraps the CCTZ library from Google. CCTZ is a modern C++11 library extending the (C++11-native) chrono type.

R> x <- nanotime("1970-01-01T00:00:00.000000001+00:00")
R> print(x)
integer64
[1] 1
R> format(x)
[1] "1970-01-01T00:00:00.000000001+00:00"
R> x <- x + 1
R> print(x)
integer64
[1] 2
R> format(x)
[1] "1970-01-01T00:00:00.000000002+00:00"
R>
R> options("width"=60)
R> v <- nanotime(Sys.time()) + 1:5
R> v
integer64
[1] 1481505724483583001 1481505724483583002
[3] 1481505724483583003 1481505724483583004
[5] 1481505724483583005
R> format(v)
[1] "2016-12-12T01:22:04.483583001+00:00"
[2] "2016-12-12T01:22:04.483583002+00:00"
[3] "2016-12-12T01:22:04.483583003+00:00"
[4] "2016-12-12T01:22:04.483583004+00:00"
[5] "2016-12-12T01:22:04.483583005+00:00"
R> 
R> z <- zoo(cbind(A=1:5, B=5:1), v)
R> options("nanotimeFormat"="%H:%M:%E*S")  ## override default
R> z
                          A B
01:47:55.812513001 1 5
01:47:55.812513002 2 4
01:47:55.812513003 3 3
01:47:55.812513004 4 2
01:47:55.812513005 5 1
R> 
R> library(data.table)
R> library(bit64)   # register some print methods for integer64
R> raw <- data.table(cbind(A=1:5, B=5:1), v)
R> fwrite(raw, file="/tmp/raw.csv")
R> cooked <- fread("/tmp/raw.csv")

The bit64 package (by Jens Oehlschlägel) supplies the integer64 type used to store the nanosecond resolution time as (positive or negative) offsets to the epoch of January 1, 1970. The RcppCCTZ package supplies the formatting and parsing routines based on the (modern C++) library CCTZ from Google.

The package is in the very early stages. Expect changes, maybe even breaking ones. But the package has some tests, and code coverage.

See the issue tickets for an up to date list of potentially desirable, possibly planned, or at least discussed items.

Once on CRAN you will be able to do

install.packages("nanotime")

Until then, or to install development versions, it can also be installed via a standard

install.packages("RcppCCTZ")   # need 0.1.0 or later
remotes::install_github("eddelbuettel/nanotime")  

Dirk Eddelbuettel

GPL (>= 2)

News

Reference manual

It appears you don't have a PDF plugin for this browser. You can click here to download the reference manual.

install.packages("nanotime")

0.2.0 by Dirk Eddelbuettel, 2 hours ago


Browse source code at https://github.com/cran/nanotime


Authors: Dirk Eddelbuettel and Leonardo Silvestri


Documentation:   PDF Manual  


GPL (>= 2) license


Imports methods, bit64, RcppCCTZ, zoo

Suggests RUnit, data.table, xts


Suggested by data.table.


See at CRAN