Closes #893. dcast handles factor levels correctly on drop=FALSE.

README.md

@@ -84,6 +84,8 @@
 
  24. `X[Y, .N]` returned the same result as `X[Y, .N, nomatch=0L]`) when `Y` contained rows that has no matches in `X`. Fixed now. Closes [#963](https:/Rdatatable/data.table/issues/963). Thanks to [this SO post](http://stackoverflow.com/q/27004002/559784) from @Alex which helped discover the bug.
 
+ 25. `data.table::dcast` handles levels in factor columns properly when `drop = FALSE`. Closes [#893](https:/Rdatatable/data.table/issues/893). Thanks to @matthieugomez for the great minimal example.
+
 #### NOTES
 
  1. Clearer explanation of what `duplicated()` does (borrowed from base). Thanks to @matthieugomez for pointing out. Closes [#872](https:/Rdatatable/data.table/issues/872).

inst/tests/tests.Rraw

@@ -5623,6 +5623,17 @@ test(1456.1, chmatch2(x1, table), as.integer(c(3,1,NA,2,5,NA)))
 test(1456.2, chmatch2(x2, table), as.integer(c(1,2,NA)))
 test(1456.3, chmatch2(x3, table), as.integer(c(NA,1,2,NA,NA)))
 
+# Fix for #893
+if ("package:reshape2" %in% search()) {
+ dt <- data.table(
+ x = factor("a", levels = c("a", "b")),
+ y = factor("b", levels = c("a", "b")),
+ z = 1
+ )
+ test(1449.1, dcast(dt, y ~ x, drop = FALSE, value.var="z"), 
+ data.table(dcast(as.data.frame(dt), y ~ x, drop = FALSE, value.var="z"), key="y"))
+}
+
 ##########################
 
 

src/fcast.c

@@ -401,11 +401,19 @@ SEXP fcast(SEXP DT, SEXP inames, SEXP mnames, SEXP vnames, SEXP fill, SEXP fill_
  for (i=0; i<mlen; i++) {
  cpy = PROTECT(allocVector(VECSXP, 1));
  SET_VECTOR_ELT(cpy, 0, VECTOR_ELT(rdt, i));
- dorder = PROTECT(cast_order(cpy, env));
- ddup = PROTECT(uniqlist(cpy, dorder));
- ddup = PROTECT(subsetVector(dorder, ddup));
- dtmp = PROTECT(subsetVector(VECTOR_ELT(cpy, 0), ddup));
- UNPROTECT(5);
+ if (isFactor(VECTOR_ELT(rdt, i))) {
+ lvls = PROTECT(getAttrib(VECTOR_ELT(rdt, i), R_LevelsSymbol));
+ dtmp = PROTECT(seq_int(length(lvls), 1));
+ setAttrib(dtmp, R_ClassSymbol, mkString("factor"));
+ setAttrib(dtmp, R_LevelsSymbol, lvls);
+ UNPROTECT(3);
+ } else {
+ dorder = PROTECT(cast_order(cpy, env));
+ ddup = PROTECT(uniqlist(cpy, dorder));
+ ddup = PROTECT(subsetVector(dorder, ddup));
+ dtmp = PROTECT(subsetVector(VECTOR_ELT(cpy, 0), ddup));
+ UNPROTECT(5);
+ }
  SET_VECTOR_ELT(rcj, i, dtmp);
  }
  rcj = PROTECT(cross_join(rcj, env)); protecti++;