:=NULL together with i is now an error

README.md

@@ -159,6 +159,8 @@ We moved from R-Forge to GitHub on 9 June 2014, including history.
       DT[,.(colB,colC,colD)]          # same
   ```
   Similarly, `by=.()` is now a shortcut for `by=list()`, for consistency with `i` and `j`.
+  
+  28. `DT[where, someCol:=NULL]` is now an error that i is provided since it makes no sense to delete a column for only a subset of rows.
           
           
 

inst/tests/tests.Rraw

@@ -4900,10 +4900,15 @@ test(1352.6, DT[,sum(b),by=.(a%%2)], DT[,sum(b),by=a%%2])
 test(1352.7, DT[,sum(b),by=.(Grp=a%%2)], DT[,sum(b),by=list(Grp=a%%2)])
 test(1352.8, DT[,sum(b),by=.(a%%2,c)], DT[,sum(b),by=list(a%%2,c)])
 
+# that :=NULL together with i is now an error
+DT = data.table(a=1:3, b=1:6)
+test(1353.1, DT[2, b:=NULL], error="When deleting columns, i should not be provided")
+test(1353.2, DT[2, c("a","b"):=list(42, NULL)], error="When deleting columns, i should not be provided")
 
 ##########################
 
 
+
 # TO DO: Add test for fixed bug #5519 - dcast.data.table returned error when a package imported data.table, but dint happen when "depends" on data.table. This is fixed (commit 1263 v1.9.3), but not sure how to add test.
 
 # TO DO: test and highlight in docs that negatives are fine and fast in forderv (ref R wish #15644)

src/assign.c

@@ -403,6 +403,7 @@ SEXP assign(SEXP dt, SEXP rows, SEXP cols, SEXP newcolnames, SEXP values, SEXP v
         else
             thisvalue = values;   // One vector applied to all columns, often NULL or NA for example
         if (TYPEOF(thisvalue)==NILSXP) {
+            if (!isNull(rows)) error("When deleting columns, i should not be provided");
             anytodelete = TRUE;
             continue;   // delete column(s) afterwards, below this loop
         }