only assign copied value in := if not null

NEWS.md

@@ -544,6 +544,8 @@
 
 51. `merge.data.table()` silently ignored the `incomparables` argument, [#2587](https://github.com/Rdatatable/data.table/issues/2587). It is now implemented and any other ignored arguments (e.g. misspellings) are now warned about. Thanks to @GBsuperman for the report and @ben-schwen for the fix.
 
+52. `DT[, c('z','x') := {x=NULL; list(2,NULL)}]` now removes column `x` as expected rather than incorrectly assigning `2` to `x` as well as `z`, [#5284](https://github.com/Rdatatable/data.table/issues/5284). The `x=NULL` is superfluous while the `list(2,NULL)` is the final value of `{}` whose items correspond to `c('z','x')`. Thanks @eutwt for the report, and @ben-schwen for the fix.
+
 ## NOTES
 
 1. New feature 29 in v1.12.4 (Oct 2019) introduced zero-copy coercion. Our thinking is that requiring you to get the type right in the case of `0` (type double) vs `0L` (type integer) is too inconvenient for you the user. So such coercions happen in `data.table` automatically without warning. Thanks to zero-copy coercion there is no speed penalty, even when calling `set()` many times in a loop, so there's no speed penalty to warn you about either. However, we believe that assigning a character value such as `"2"` into an integer column is more likely to be a user mistake that you would like to be warned about. The type difference (character vs integer) may be the only clue that you have selected the wrong column, or typed the wrong variable to be assigned to that column. For this reason we view character to numeric-like coercion differently and will warn about it. If it is correct, then the warning is intended to nudge you to wrap the RHS with `as.<type>()` so that it is clear to readers of your code that a coercion from character to that type is intended. For example :

R/data.table.R

@@ -1377,7 +1377,7 @@ replace_dot_alias = function(e) {
       } else if (address(jval) == address(SDenv$.SD)) {
         jval = copy(jval)
       } else if ( length(jcpy <- which(vapply_1c(jval, address) %chin% vapply_1c(SDenv, address))) ) {
-        for (jidx in jcpy) jval[[jidx]] = copy(jval[[jidx]])
+        for (jidx in jcpy) { if(!is.null(jval[[jidx]])) jval[[jidx]] = copy(jval[[jidx]]) }
       } else if (jsub %iscall% 'get') {
         jval = copy(jval) # fix for #1212
       }

inst/tests/tests.Rraw

@@ -18765,3 +18765,7 @@ test(2234.7, DT[, min(.SD), by=c(.I)], data.table(I=1L:5L, V1=c(1L, 2L, 3L, 2L,
 test(2234.8, DT[, min(.SD), by=.I%%2L], error="by.*contains .I.*supported")  # would be nice to support in future; i.e. by odd/even rows, and by=(.I+1L)%/%2L for pairs of rows; i.e. any expression of .I
 test(2234.9, DT[, min(.SD), by=somefun(.I)], error="by.*contains .I.*supported")
 
+# copying values of j could lead to recycling if j is a list containing NULL #5284
+DT = data.table(x = 1)
+test(2235.1, copy(DT)[, c("z", "x") := {x = NULL; list(2, NULL)}], data.table(z = 2))
+test(2235.2, copy(DT)[, c("z", "x") := {list(2, NULL)}], data.table(z = 2))