fix #6898, disable checking timezone in between()

Author: Bill Evans

NEWS.md

@@ -10,6 +10,8 @@
 
 3. `print.data.table()` now shows column classes at the bottom of large tables when `class=TRUE` and `col.names="auto"` (default) for tables with more than 20 rows, [#6902](https://github.com/Rdatatable/data.table/issues/6902). This follows the same behavior as column names at the bottom, making it easier to see column types for large tables without scrolling back to the top. Thanks to @TimTaylor for the suggestion and @Mukulyadav2004 for the PR.
 
+4. `between()` gains the argument `ignore_tzone=FALSE`. Normally, a difference in time zone between `lower` and `upper` will produce an error, and a difference in time zone between `x` and either of the others will produce a message. Setting `ignore_tzone=TRUE` bypasses the checks, allowing both comparisons to proceed without error or message about time zones.
+
 ## BUG FIXES
 
 1. Custom binary operators from the `lubridate` package now work with objects of class `IDate` as with a `Date` subclass, [#6839](https://github.com/Rdatatable/data.table/issues/6839). Thanks @emallickhossain for the report and @aitap for the fix.

R/between.R

@@ -1,5 +1,5 @@
 # is x[i] in between lower[i] and upper[i] ?
-between = function(x, lower, upper, incbounds=TRUE, NAbounds=TRUE, check=FALSE) {
+between = function(x, lower, upper, incbounds=TRUE, NAbounds=TRUE, check=FALSE, ignore_tzone=FALSE) {
   if (is.logical(x)) stopf("between has been passed an argument x of type logical")
   if (is.logical(lower)) lower = as.integer(lower)   # typically NA (which is logical type)
   if (is.logical(upper)) upper = as.integer(upper)   # typically NA (which is logical type)
@@ -16,7 +16,7 @@ between = function(x, lower, upper, incbounds=TRUE, NAbounds=TRUE, check=FALSE)
     stopifnot(is.px(x), is.px(lower), is.px(upper)) # nocov # internal
   }
   # POSIX check timezone match
-  if (is.px(x) && is.px(lower) && is.px(upper)) {
+  if (!ignore_tzone && is.px(x) && is.px(lower) && is.px(upper)) {
     tzs = sapply(list(x,lower,upper), function(x) {
       attr(x, "tzone", exact=TRUE) %||% ""
     })

inst/tests/tests.Rraw

@@ -21141,3 +21141,13 @@ dt = data.table(id = 1:25)
 test(2314.1, any(grepl("<int>", tail(capture.output(print(dt, class = TRUE)), 2))), TRUE)
 # Test that class=TRUE with col.names="top" doesn't show classes at bottom
 test(2314.2, !any(grepl("<int>", tail(capture.output(print(dt, class = TRUE, col.names = "top")), 2))), TRUE)
+
+# Test that tzone behavior stops warning with ignore_tzone=TRUE
+tm1 = .POSIXct(1, "UTC")
+dt1 = data.table(a = .POSIXct(0), b = .POSIXct(2, "UTC"))
+test(2315.1, dt1[, between(tm1, a, b)], error = "different tzone attributes")
+test(2315.2, dt1[, between(tm1, a, b, ignore_tzone=TRUE)])
+tm2 <- .POSIXct(1)
+dt2 = data.table(a = .POSIXct(0, "UTC"), b = .POSIXct(2, "UTC"))
+test(2315.3, dt2[, between(tm2, a, b)], message = "mismatched tzone attributes")
+test(2315.4, dt2[, between(tm2, a, b)])

man/between.Rd

@@ -35,6 +35,8 @@ interpreted as \code{lower} and \code{y[[2]]} as \code{upper}.}
 It is set to \code{TRUE} by default for infix notations.}
 \item{NAbounds}{ If \code{lower} (\code{upper}) contains an \code{NA} what should \code{lower<=x} (\code{x<=upper}) return? By default \code{TRUE} so that a missing bound is interpreted as unlimited. }
 \item{check}{ Produce error if \code{any(lower>upper)}? \code{FALSE} by default for efficiency, in particular type \code{character}. }
+\item{ignore_tzone}{ \code{TRUE} means skip timezone checks among
+\code{x}, \code{lower}, and \code{upper}. }
 }
 \details{