These function use the Boost Date_Time library to parse
datetimes (and dates) from strings, integers, factors or even numeric values
(which are cast to strings internally). They return a vector of
POSIXct objects (or Date objects in the case of anydate).
POSIXct objects represent dates and time as (possibly
fractional) seconds since the ‘epoch’ of January 1, 1970.
A timezone can be set, if none is supplied ‘UTC’ is set.
anytime(x, tz = getTZ(), asUTC = FALSE,
useR = getOption("anytimeUseRConversions", FALSE),
oldHeuristic = getOption("anytimeOldHeuristic", FALSE),
calcUnique = FALSE)anydate(x, tz = getTZ(), asUTC = FALSE,
useR = getOption("anytimeUseRConversions", FALSE), calcUnique = FALSE)
utctime(x, tz = getTZ(), useR = getOption("anytimeUseRConversions", FALSE),
oldHeuristic = getOption("anytimeOldHeuristic", FALSE),
calcUnique = FALSE)
utcdate(x, tz = getTZ(), useR = getOption("anytimeUseRConversions", FALSE),
calcUnique = FALSE)
A vector of POSIXct elements, or, in the case of anydate,
a vector of Date objects.
A vector of type character, integer or numeric with date(time) expressions to be parsed and converted.
A string with the timezone, defaults to the result of the (internal)
getTZ function if unset. The getTZ function returns the timezone
values stored in local package environment, and set at package load time. Also
note that this argument applies to the output: the returned object will
have this timezone set. The timezone is not used for the parsing which
will always be to localtime, or to UTC is the asUTC variable is set (as
it is in the related functions utctime and utcdate).
So one can think of the argument as ‘shift parsed time object to this
timezone’. This is similar to what format() in base R does, but our
return value is still a POSIXt object instead of a character value.
A logical value indicating if parsing should be to UTC; default is false implying localtime.
A logical value indicating if conversion should be done via code
from R (via Rcpp::Function) instead of the default Boost routines. The
default value is the value of the option anytimeUseRConversions with a
fallback of FALSE if the option is unset. In other words, this will
be false by default but can be set to true via an option.
A logical value to enable behaviour as in version 0.2.2 or earlier:
interpret a numeric or integer value that could be seen as a YYYYMMDD as a date. If
the default value FALSE is seen, then numeric values are used as offsets
dates (in anydate or utcdate), and as second offsets for datetimes
otherwise. A default value can also be set via the anytimeOldHeuristic option.
A logical value with a default value of FALSE that tells the
function to perform the anytime() or anydate() calculation only once for
each unique value in the x vector. It results in no difference in inputs or
outputs, but can result in a significant speed increases for long vectors where each
timestamp appears more than once. However, it will result in a slight slow down for
input vectors where each timestamp appears only once.
By default, the (internal) conversion to (fractional) seconds since the epoch is relative to the locatime of this system, and therefore not completely independent of the settings of the local system. This is to strike a balance between ease of use and functionality. A more-full featured conversion could be possibly be added with support for arbitrary reference times, but this is (at least) currently outside the scope of this package. See the RcppCCTZ package which offers some timezone-shifting and differencing functionality. As of version 0.0.5 one can also parse relative to UTC avoiding the localtime issue,
Times and timezones can be tricky. This package offers a heuristic approach, it is likely that some input formats may not be parsed, or worse, be parsed incorrectly. This is not quite a Bobby Tables situation but care must always be taken with user-supplied input.
The Boost Date_Time library cannot parse single digit months or days. So while ‘2016/09/02’ works (as expected), ‘2016/9/2’ will not. Other non-standard formats may also fail.
There is a known issue (discussed at length in issue ticket 5) where Australian times are off by an hour. This seems to affect only Windows, not Linux.
When given a vector, R will coerce it to the type of the first
element. Should that be NA, surprising things can
happen: c(NA, Sys.Date()) forces both values to
numeric and the date will not be parsed correctly (as its
integer value becomes numeric before our code sees it). On the
other hand, c(Sys.Date(), NA) works as expected parsing as
type Date with one missing value. See
issue
ticket 11 for more.
Another known issue concerns conversion when the timezone is set
to ‘Europe/London’, see GitHub issue tickets
36.
51.
59. and
86. As
pointed out in the comment in that last one, the
Sys.timezone manual page suggests several
alternatives to using ‘Europe/London’ such as ‘GB’.
Yet another known issue arises on Windows due to designs in the
Boost library. While we can set the TZ library variable,
Boost actually does not consult it but rather relies only
on the (Windows) tool tzutil. This means that default
behaviour should be as expected: dates and/or times are parsed to
the local settings. But testing different TZ values (or
more precisely, changes via the (unexported) helper function
setTZ function as we cache TZ) will only influence
the behaviour on Unix or Unix-alike operating systems and not on
Windows. See the discussion at
issue
ticket 96 for more. In short, the recommendation for Windows
user is to also set useR=TRUE when setting a timezone
argument.
On Windows systems, accessing the isdst flag on dates or times
before January 1, 1970, can lead to a crash. Therefore, the lookup of this
value has been disabled for those dates and times, which could therefore be
off by an hour (the common value that needs to be corrected).
It should not affect dates, but may affect datetime objects.
Up until version 0.2.2, numeric input smaller than an internal cutoff value
was interpreted as a date, even if anytime() was called. While
convenient, it is also inconsistent as we otherwise take numeric values to
be offsets to the epoch. Newer version are consistent: for anydate, a
value is taken as date offset relative to the epoch (of January 1, 1970).
For anytime, it is taken as seconds offset. So anytime(60)
is one minute past the epoch, and anydate(60) is sixty days past it.
The old behaviour can be enabled by setting the oldHeuristic argument to
anytime (and utctime) to TRUE. Additionally, the default
value can be set via getOption("anytimeOldHeuristic") which can be set
to TRUE in startup file. Note that all other inputs such character,
factor or ordered are not affected.
As of version 0.3.10, a conversion from character resulting in a NA
will lead to a warning being emitted. At most one warning per call is given:
should numerous unparseable values be present on input, only one warning will
be show. R offers mechanism to either suppress warnings, or convert them
to errors as described in the help page for options() under the entry
for warn.
Dirk Eddelbuettel
A number of fixed formats are tried in succession. These include
the standard ISO format ‘YYYY-MM-DD HH:MM:SS’ as well as
different local variants including several forms popular in the
United States. Two-digits years and clearly ambigous formats such
as ‘03/04/05’ are ignored. In the case of parsing failure
a NA value is returned.
Fractional seconds are supported as well. As R itself only supports microseconds, the Boost compile-time option for nano-second resolution has not been enabled.
This StackOverflow answer provided the initial idea: https://stackoverflow.com/a/3787188/143305.
anytime-package, getFormats
## See the source code for a full list of formats, and the
## or the reference in help('anytime-package') for details
times <- c("2004-03-21 12:45:33.123456",
"2004/03/21 12:45:33.123456",
"20040321 124533.123456",
"03/21/2004 12:45:33.123456",
"03-21-2004 12:45:33.123456",
"2004-03-21",
"20040321",
"03/21/2004",
"03-21-2004",
"20010101")
anytime(times)
anydate(times)
utctime(times)
utcdate(times)
## show effect of tz argument
anytime("2001-02-03 04:05:06")
## adjust parsed time to given TZ argument
anytime("2001-02-03 04:05:06", tz="America/Los_Angeles")
## somewhat equvalent base R functionality
format(anytime("2001-02-03 04:05:06"), tz="America/Los_Angeles")
Run the code above in your browser using DataLab