Merge branch 'master' into i64remapfix

Author: badasahog

.github/workflows/performance-tests.yml

@@ -20,4 +20,4 @@ jobs:
       GITHUB_PAT: ${{ secrets.GITHUB_TOKEN }}
       repo_token: ${{ secrets.GITHUB_TOKEN }}
     steps:
-      - uses: Anirban166/[email protected].1
+      - uses: Anirban166/[email protected].2

NEWS.md

@@ -16,6 +16,8 @@
 
 5. `as.data.table()` is slightly more efficient at converting arrays to data.tables, [#7019](https://github.com/Rdatatable/data.table/pull/7019). Thanks @eliocamp.
 
+6. `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.
@@ -34,7 +36,9 @@
 
 8. `fread()` no longer warns on certain systems on R 4.5.0+ where the file owner can't be resolved, [#6918](https://github.com/Rdatatable/data.table/issues/6918). Thanks @ProfFancyPants for the report and PR.
 
-9. Out of sample type bumps now respect `integer64=` selection, [#7032](https://github.com/Rdatatable/data.table/pull/7032)
+9. Joins to extended data.frames, e.g. `x[i, col := x.col1 + i.col2]` where `i` is a `tbl`, can use the `x.` and `i.` prefix forms, [#6998](https://github.com/Rdatatable/data.table/issues/6998). Thanks @MichaelChirico for the bug and PR.
+
+10. Out of sample type bumps now respect `integer64=` selection, [#7032](https://github.com/Rdatatable/data.table/pull/7032)
 
 ### NOTES
 

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)
@@ -16,15 +16,12 @@ 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)) {
-    tzs = sapply(list(x,lower,upper), function(x) {
-      attr(x, "tzone", exact=TRUE) %||% ""
-    })
+  if (!ignore_tzone && is.px(x) && is.px(lower) && is.px(upper)) {
+    tzs = vapply_1c(list(x, lower, upper), function(x) attr(x, "tzone", exact=TRUE) %||% "")
     # lower/upper should be more tightly linked than x/lower, so error
     #   if the former don't match but only inform if they latter don't
     if (tzs[2L]!=tzs[3L]) {
       stopf("'between' lower= and upper= are both POSIXct but have different tzone attributes: %s. Please align their time zones.", brackify(tzs[2:3], quote=TRUE))
-      # otherwise the check in between.c that lower<=upper can (correctly) fail for this reason
     }
     if (tzs[1L]!=tzs[2L]) {
       messagef("'between' arguments are all POSIXct but have mismatched tzone attributes: %s. The UTC times will be compared.", brackify(tzs, quote=TRUE))

R/data.table.R

@@ -677,10 +677,12 @@ replace_dot_alias = function(e) {
     }
     ansvals = chmatch(ansvars, nx)
   } else {
-    if (is.data.table(i)) {
+    if (is.data.frame(i)) {
       idotprefix = paste0("i.", names_i)
       xdotprefix = paste0("x.", names_x)
-    } else idotprefix = xdotprefix = character(0L)
+    } else {
+      idotprefix = xdotprefix = character(0L)
+    }
 
     # j was substituted before dealing with i so that := can set allow.cartesian=FALSE (#800) (used above in i logic)
     if (is.null(jsub)) return(NULL)

inst/tests/tests.Rraw

@@ -21133,3 +21133,27 @@ test(2315.1, tail(DT[order(i), i], 2L), 1:2)
 # wider range of numbers needed for further coverage
 DT[1L, i := 1000L]
 test(2315.2, tail(DT[order(i), i], 2L), c(1L, 1000L))
+
+# issue #6898, test that tzone behavior changes with ignore_tzone=TRUE
+tms = list(.POSIXct(1), .POSIXct(1.0, "UTC"))
+test(2316.1, between(tms[[1]], tms[[1L]], tms[[2L]]), error = "different tzone attributes")
+test(2316.2, between(tms[[1]], tms[[1L]], tms[[2L]], ignore_tzone=TRUE))
+test(2316.3, between(tms[[1]], tms[[2L]], tms[[2L]]), message = "mismatched tzone attributes")
+test(2316.4, between(tms[[1]], tms[[2L]], tms[[2L]], ignore_tzone=TRUE))
+
+# tbl in i still allows 'i.' prefix reference for update join, #6998
+DT1 = data.table(a=1, b=2)
+DT2 = data.table(a=1, c=3)
+DF1 = data.frame(a=1, d=4)
+DF2 = data.frame(a=1, e=5)
+class(DF2) = c("tbl_df", "tbl", "data.frame")
+
+test(2317.1, DT1[DT2, on='a', c := i.c]$c, 3)
+test(2317.2, DT1[DT2, on='a', c2 := x.a + i.c]$c2, 4)
+test(2317.3, DT1[DT2, on='a', .(c = x.a + i.c)]$c, 4)
+test(2317.4, DT1[DF1, on='a', d := i.d]$d, 4)
+test(2317.5, DT1[DF1, on='a', d2 := x.a + i.d]$d2, 5)
+test(2317.6, DT1[DF1, on='a', .(d = x.a + i.d)]$d, 5)
+test(2317.7, DT1[DF2, on='a', e := i.e]$e, 5)
+test(2317.8, DT1[DF2, on='a', e2 := x.a + i.e]$e2, 6)
+test(2317.9, DT1[DF2, on='a', .(e = x.a + i.e)]$e, 6)

man/between.Rd

@@ -16,7 +16,7 @@ This can be changed by setting \code{NAbounds} to \code{NA}.
 the intervals provided in \code{lower,upper}.
 }
 \usage{
-between(x, lower, upper, incbounds=TRUE, NAbounds=TRUE, check=FALSE)
+between(x, lower, upper, incbounds=TRUE, NAbounds=TRUE, check=FALSE, ignore_tzone=FALSE)
 x \%between\% y
 
 inrange(x, lower, upper, incbounds=TRUE)
@@ -35,6 +35,7 @@ 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{