set automatically allocates new column slots if needed

NEWS.md

@@ -455,6 +455,8 @@
 
 49. `fintersect(..., all=TRUE)` and `fsetdiff(..., all=TRUE)` could return incorrect results when the inputs had columns named `x` and `y`, [#5255](https://github.com/Rdatatable/data.table/issues/5255). Thanks @Fpadt for the report, and @ben-schwen for the fix.
 
+50. `set()` now automatically pre-allocates new column slots if needed, similar to what `:=` already does, [#496](https://github.com/Rdatatable/data.table/issues/496) [#4100](https://github.com/Rdatatable/data.table/issues/4100). Thanks to Huashan Chen and Benjamin Tyner for the report and Benjamin Schwendinger 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

@@ -2707,6 +2707,22 @@ setcolorder = function(x, neworder=key(x), before=NULL, after=NULL)  # before/af
 
 set = function(x,i=NULL,j,value)  # low overhead, loopable
 {
+  if (is.character(j) && truelength(x) < ncol(x)+length(j)) {
+    newnames = NULL
+    names_x = names(x)
+    m = chmatch(j, names_x)
+    if (anyNA(m)) {
+      newnames = j[is.na(m)]
+    }
+    if (truelength(x) < ncol(x)+length(newnames)) {
+      n = length(newnames) + eval(getOption("datatable.alloccol"))
+      name = substitute(x)
+      x = .Call(Calloccolwrapper, x, eval(n), getOption("datatable.verbose"))
+      if (is.name(name)) {
+        assign(as.character(name),x,parent.frame(),inherits=TRUE)
+      }
+    }
+  }
   .Call(Cassign,x,i,j,NULL,value)
   invisible(x)
 }

inst/tests/tests.Rraw

@@ -1265,7 +1265,8 @@ test(416, DT[J(2)], error="the columns to join by must be specified using")
 
 # Test shallow copy verbose message from := adding a column, and (TO DO) only when X is NAMED.
 DT = data.table(a=1:3,b=4:6)
-test(417, alloc.col(DT,3,verbose=TRUE), DT, output="Attempt to reduce allocation from.*to 5 ignored. Can only increase allocation via shallow copy")
+test(417.1, alloc.col(DT,3,verbose=TRUE), DT, output="Attempt to reduce allocation from.*to 5 ignored. Can only increase allocation via shallow copy")
+test(417.2, alloc.col(DT,3,verbose=1L), error="verbose must be TRUE or FALSE")
 options(datatable.alloccol=1L)
 DT = data.table(a=1:3,b=4:6)
 options(datatable.alloccol=1024L)
@@ -14784,7 +14785,7 @@ test(2016.1, name, "DT")
 test(2016.2, DT, data.table(a=1:3))
 test(2016.3, DT[2,a:=4L], data.table(a=INT(1,4,3)))        # no error for := when existing column
 test(2016.4, set(DT,3L,1L,5L), data.table(a=INT(1,4,5)))   # no error for set() when existing column
-test(2016.5, set(DT,2L,"newCol",5L), error="either been loaded from disk.*or constructed manually.*Please run setDT.*setalloccol.*on it first")   # just set()
+test(2016.5, set(DT,2L,"newCol",5L), data.table(a=INT(1,4,5), newCol=INT(NA,5L,NA)))   # works since set overallocates
 test(2016.6, DT[2,newCol:=6L], data.table(a=INT(1,4,5), newCol=INT(NA,6L,NA)))  # := ok (it changes DT in caller)
 unlink(tt)
 
@@ -18382,3 +18383,10 @@ DT = data.table(x=c(1,2,2,2), y=LETTERS[c(1,2,2,3)])
 test(2227.1, fintersect(DT, DT, all=TRUE), DT)
 test(2227.2, fsetdiff(DT, DT, all=TRUE), DT[0])
 
+# re-overallocate in set if quota is reached #496 #4100
+opt = options(datatable.alloccol=1L)
+n = getOption("datatable.alloccol")
+DT = data.table()
+for (i in seq(2*n)) set(DT, j = paste0("V",i), value = i)
+test(2228.1, ncol(DT), 2L*n)
+options(opt)

src/assign.c

@@ -459,7 +459,8 @@ SEXP assign(SEXP dt, SEXP rows, SEXP cols, SEXP newcolnames, SEXP values)
     oldtncol = TRUELENGTH(dt);   // TO DO: oldtncol can be just called tl now, as we won't realloc here any more.
 
     if (oldtncol<oldncol) {
-      if (oldtncol==0) error(_("This data.table has either been loaded from disk (e.g. using readRDS()/load()) or constructed manually (e.g. using structure()). Please run setDT() or setalloccol() on it first (to pre-allocate space for new columns) before assigning by reference to it."));   // #2996
+      // oldtncol=0 and oldtncol < oldncol can't occur longer because all functions which use assign (set and :=) pre-allocate column slots if needed
+      //if (oldtncol==0) error(_("This data.table has either been loaded from disk (e.g. using readRDS()/load()) or constructed manually (e.g. using structure()). Please run setDT() or setalloccol() on it first (to pre-allocate space for new columns) before assigning by reference to it."));   // #2996
       error(_("Internal error: oldtncol(%d) < oldncol(%d). Please report to data.table issue tracker, including result of sessionInfo()."), oldtncol, oldncol); // # nocov
     }
     if (oldtncol>oldncol+10000L) warning(_("truelength (%d) is greater than 10,000 items over-allocated (length = %d). See ?truelength. If you didn't set the datatable.alloccol option very large, please report to data.table issue tracker including the result of sessionInfo()."),oldtncol, oldncol);