Progress deprecation of with=FALSE + :=

Author: MichaelChirico

NEWS.md

@@ -137,6 +137,8 @@ rowwiseDT(
 
 6. `measurev()` was implemented and documented in v1.15.0, for use within `melt()`, and it is now exported (dependent packages can now use without a NOTE from CRAN check).
 
+7. Deprecation of `:=` and `with=FALSE` has been upgraded from warning (since v1.15.0) to error. Long ago (before 2014), this was needed when, e.g., assigning to a vector of column names defined outside the table, but `with=FALSE` is no longer needed to do so: `DT[, (cols) := ...]` works fine.
+
 # data.table [v1.16.2](https://github.com/Rdatatable/data.table/milestone/35) (9 October 2024)
 
 ## BUG FIXES

R/data.table.R

@@ -685,19 +685,11 @@ replace_dot_alias = function(e) {
     # 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)
 
-    if (!with && jsub %iscall% ":=") {
-      # TODO: make these both errors (or single long error in both cases) in next release.
-      # i.e. using with=FALSE together with := at all will become an error. Eventually with will be removed.
-      if (is.null(names(jsub)) && is.name(jsub[[2L]])) {
-        warningf("with=FALSE together with := was deprecated in v1.9.4 released Oct 2014. Please wrap the LHS of := with parentheses; e.g., DT[,(myVar):=sum(b),by=a] to assign to column name(s) held in variable myVar. See ?':=' for other examples. As warned in 2014, this is now a warning.")
-        jsub[[2L]] = eval(jsub[[2L]], parent.frame(), parent.frame())
-      } else {
-        warningf("with=FALSE ignored, it isn't needed when using :=. See ?':=' for examples.")
-      }
-      with = TRUE
-    }
-
     if (!with) {
+      if (jsub %iscall% ":=") {
+        # TODO(>=1.18.0): Simplify this error
+        stopf("with=FALSE together with := was deprecated in v1.9.4 released Oct 2014; this has been warning since v1.15.0. Please wrap the LHS of := with parentheses; e.g., DT[,(myVar):=sum(b),by=a] to assign to column name(s) held in variable myVar. See ?':=' for other examples.")
+      }
       # missingby was already checked above before dealing with i
       if (jsub %iscall% c("!", "-") && length(jsub)==2L) {  # length 2 to only match unary, #2109
         notj = TRUE

inst/tests/tests.Rraw

@@ -4658,7 +4658,7 @@ test(1241, DT[order(x,-y)],       # optimized to forder()
 
 DT = data.table(a=1:3, b=4:6)
 myCol = "a"
-test(1242.1, DT[2,myCol:=6L,with=FALSE], data.table(a=INT(1,6,3), b=4:6), warning="with=FALSE together with := was deprecated in v1.9.4 released Oct 2014. Please")
+test(1242.1, DT[2,myCol:=6L,with=FALSE], error="with=FALSE together with := was deprecated in v1.9.4)
 test(1242.2, DT[2,(myCol):=7L], data.table(a=INT(1,7,3), b=4:6))
 
 # consistency of output type of mult, #340
@@ -13935,8 +13935,7 @@ test(1967.42, x[3, rollends = rep(TRUE, 10L)], error = 'rollends must be length
 test(1967.43, x[ , ..], error = 'symbol .. is invalid')
 test(1967.44, x[NULL], data.table(NULL))
 test(1967.45, x[ , NULL], NULL)
-test(1967.46, x[ , 'b' := 6:10, with = FALSE],
-     data.table(a = 1:5, b = 6:10), warning = 'with=FALSE ignored')
+test(1967.46, x[ , 'b' := 6:10, with=FALSE], error='with=FALSE ignored')
 test(1967.47, x[ , -1L, with = FALSE], data.table(b = 6:10))
 test(1967.48, x[ , b, .SDcols = 'a'], 6:10,
      warning = "This j doesn't use .SD")