Merge branch 'master' into constnessAndReadOnlyness

Author: MichaelChirico

NAMESPACE

@@ -139,7 +139,6 @@ S3method(split, data.table)
 export(dcast, melt)
 S3method(dcast, data.table)
 S3method(melt, data.table)
-S3method(melt, default)
 
 # exported for historical reasons -- if reshape2 is higher on search path,
 # dcast(DT) will not dispatch since reshape2::dcast is not generic. So users

NEWS.md

@@ -46,7 +46,7 @@
 
 9. `foverlaps()` no longer crashes due to out-of-bounds access to list and integer vectors when `y` has no rows or the non-range part of the join fails, [#7597](https://github.com/Rdatatable/data.table/issues/7597). Thanks to @nextpagesoft for the report and @aitap for the fix.
 
-10. The dynamic library now exports only `R_init_data_table`, preventing symbol name conflicts like `hash_create` with PostgreSQL, [#7605](https://github.com/Rdatatable/data.table/issues/7605). Thanks to @ced75 for the report and @aitap for the fix
+10. The dynamic library now exports only `R_init_data_table`, preventing symbol name conflicts like `hash_create` with PostgreSQL, [#7605](https://github.com/Rdatatable/data.table/issues/7605). Thanks to @ced75 for the report and @aitap for the fix.
 
 ### Notes
 
@@ -60,6 +60,8 @@
 
 5. The data.table test suite is a bit more robust to lacking UTF-8 support via a new `requires_utf8` argument to `test()` to skip tests when UTF-8 support is not available, [#7336](https://github.com/Rdatatable/data.table/issues/7336). Thanks @MichaelChirico for the suggestion and @ben-schwen for the implementation.
 
+6. `melt()` and `dcast()` no longer provide nudges when receiving incompatible inputs (e.g. data.frames). As of now, we only define methods for `data.table` inputs.
+
 ## data.table [v1.18.0](https://github.com/Rdatatable/data.table/milestone/37?closed=1)  23 December 2025
 
 ### BREAKING CHANGE

R/fcast.R

@@ -12,9 +12,6 @@ dcast = function(
   data, formula, fun.aggregate = NULL, ..., margins = NULL,
   subset = NULL, fill = NULL, value.var = guess(data)
 ) {
-  # TODO(>=1.19.0): Remove this, just let dispatch to 'default' method fail.
-  if (!is.data.table(data))
-    stopf("The %1$s generic in data.table has been passed a %2$s, but data.table::%1$s currently only has a method for data.tables. Please confirm your input is a data.table, with setDT(%3$s) or as.data.table(%3$s). If you intend to use a method from reshape2, try installing that package first, but do note that reshape2 is superseded and is no longer actively developed.", "dcast", class1(data), deparse(substitute(data))) # nocov
   UseMethod("dcast", data)
 }
 

R/fmelt.R

@@ -7,11 +7,6 @@ melt = function(data, ..., na.rm = FALSE, value.name = "value") {
   UseMethod("melt", data)
 }
 
-# TODO(>=1.19.0): Remove this, just let dispatch to 'default' method fail.
-melt.default = function(data, ..., na.rm = FALSE, value.name = "value") {
-  stopf("The %1$s generic in data.table has been passed a %2$s and will attempt to redirect to the relevant reshape2 method; please note that reshape2 is superseded and is no longer actively developed, and this redirection is now deprecated. To continue using melt methods from reshape2 while both packages are attached, e.g. melt.list, you can prepend the namespace, i.e. reshape2::%1$s(%3$s). In the next version, this warning will become an error.", "melt", class1(data), deparse(substitute(data))) # nocov
-}
-
 patterns = function(..., cols=character(0L), ignore.case=FALSE, perl=FALSE, fixed=FALSE, useBytes=FALSE) {
   # if ... has no names, names(list(...)) will be "";
   #   this assures they'll be NULL instead

src/assign.c

@@ -403,7 +403,7 @@ SEXP assign(SEXP dt, SEXP rows, SEXP cols, SEXP newcolnames, SEXP values)
       error(_("i is type '%s'. Must be integer, or numeric is coerced with warning. If i is a logical subset, simply wrap with which(), and take the which() outside the loop if possible for efficiency."), type2char(TYPEOF(rows)));
     targetlen = length(rows);
     numToDo = 0;
-    const int *rowsd = INTEGER(rows);
+    const int *rowsd = INTEGER_RO(rows);
     for (int i=0; i<targetlen; ++i) {
       if ((rowsd[i]<0 && rowsd[i]!=NA_INTEGER) || rowsd[i]>nrow)
         error(_("i[%d] is %d which is out of range [1,nrow=%d]"), i+1, rowsd[i], nrow);  // set() reaches here (test 2005.2); := reaches the same error in subset.c first
@@ -660,7 +660,7 @@ SEXP assign(SEXP dt, SEXP rows, SEXP cols, SEXP newcolnames, SEXP values)
   if (ndelete) {
     // delete any columns assigned NULL (there was a 'continue' earlier in loop above)
     int *tt = (int *)R_alloc(ndelete, sizeof(*tt));
-    const int *colsd=INTEGER(cols), ncols=length(cols), ndt=length(dt);
+    const int *colsd=INTEGER_RO(cols), ncols=length(cols), ndt=length(dt);
     for (int i=0, k=0; i<ncols; ++i) {   // find which ones to delete and put them in tt
       // Aside: a new column being assigned NULL (something odd to do) would have been warned above, added above, and now deleted. Just
       //        easier to code it this way; e.g. so that other columns may be added or removed ok by the same query.
@@ -743,7 +743,7 @@ const char *memrecycle(const SEXP target, const SEXP where, const int start, con
         // allow assigning level numbers to factor columns; test 425, 426, 429 and 1945
         const int nlevel = length(getAttrib(target, R_LevelsSymbol));
         if (isInteger(source)) {
-          const int *sd = INTEGER(source);
+          const int *sd = INTEGER_RO(source);
           for (int i=0; i<slen; ++i) {
             const int val = sd[i+soff];
             if ((val<1 && val!=NA_INTEGER) || val>nlevel) {
@@ -754,7 +754,7 @@ const char *memrecycle(const SEXP target, const SEXP where, const int start, con
             }
           }
         } else {
-          const double *sd = REAL(source);
+          const double *sd = REAL_RO(source);
           for (int i=0; i<slen; ++i) {
             const double val = sd[i+soff];
             // Since nlevel is an int, val < 1 || val > nlevel will deflect UB guarded against in PR #5832
@@ -800,7 +800,7 @@ const char *memrecycle(const SEXP target, const SEXP where, const int start, con
         const int nSource = length(source);
         int *newSourceD = INTEGER(newSource);
         if (sourceIsFactor) {
-          const int *sourceD = INTEGER(source);
+          const int *sourceD = INTEGER_RO(source);
           for (int i=0; i<nSource; ++i) {  // convert source integers to refer to target levels
             const int val = sourceD[i];
             newSourceD[i] = val==NA_INTEGER ? NA_INTEGER : -hash_lookup(marks, sourceLevelsD[val-1], 0); // retains NA factor levels here via TL(NA_STRING); e.g. ordered factor

src/bmerge.c

@@ -357,8 +357,8 @@ void bmerge_r(int xlowIn, int xuppIn, int ilowIn, int iuppIn, int col, int thisg
 
   switch (TYPEOF(xc)) {
   case LGLSXP : case INTSXP : {  // including factors
-    const int *icv = isDataCol ? INTEGER(ic) : NULL;
-    const int *xcv = INTEGER(xc);
+    const int *icv = isDataCol ? INTEGER_RO(ic) : NULL;
+    const int *xcv = INTEGER_RO(xc);
     const int ival = isDataCol ? icv[ir] : thisgrp;
     #define ISNAT(x) ((x)==NA_INTEGER)
     #define WRAP(x) (x)  // wrap not needed for int
@@ -389,8 +389,8 @@ void bmerge_r(int xlowIn, int xuppIn, int ilowIn, int iuppIn, int col, int thisg
       #define WRAP(x) (x)
       DO(const int64_t xval=xcv[XIND(mid)], xval<ival, xval>ival, int64_t, ival-xcv[XIND(xlow)], xcv[XIND(xupp)]-ival, ival)
     } else {
-      const double *icv = REAL(ic);
-      const double *xcv = REAL(xc);
+      const double *icv = REAL_RO(ic);
+      const double *xcv = REAL_RO(xc);
       const double ival = icv[ir];
       const uint64_t ivalt = dtwiddle(ival); // TO: remove dtwiddle by dealing with NA, NaN, -Inf, +Inf up front
       #undef ISNAT

src/data.table.h

@@ -329,6 +329,7 @@ SEXP fitsInInt64R(SEXP x);
 bool allNA(SEXP x, bool errorForBadType);
 SEXP colnamesInt(SEXP x, SEXP cols, SEXP check_dups, SEXP skip_absent);
 bool INHERITS(SEXP x, SEXP char_);
+void copyVectorElements(SEXP dst, SEXP src, R_xlen_t n, bool deep_copy, const char *caller);
 SEXP copyAsPlain(SEXP x);
 void copySharedColumns(SEXP x);
 SEXP lock(SEXP x);

src/dogroups.c

@@ -526,25 +526,7 @@ SEXP growVector(SEXP x, const R_len_t newlen)
     UNPROTECT(1);
     return newx;
   }
-  switch (TYPEOF(x)) {
-  case RAWSXP:  memcpy(RAW(newx),     RAW_RO(x),     len*RTYPE_SIZEOF(x)); break;
-  case LGLSXP:  memcpy(LOGICAL(newx), LOGICAL_RO(x), len*RTYPE_SIZEOF(x)); break;
-  case INTSXP:  memcpy(INTEGER(newx), INTEGER_RO(x), len*RTYPE_SIZEOF(x)); break;
-  case REALSXP: memcpy(REAL(newx),    REAL_RO(x),    len*RTYPE_SIZEOF(x)); break;
-  case CPLXSXP: memcpy(COMPLEX(newx), COMPLEX_RO(x), len*RTYPE_SIZEOF(x)); break;
-  case STRSXP : {
-    const SEXP *xd = SEXPPTR_RO(x);
-    for (int i=0; i<len; ++i)
-      SET_STRING_ELT(newx, i, xd[i]);
-  } break;
-  case VECSXP : {
-    const SEXP *xd = SEXPPTR_RO(x);
-    for (int i=0; i<len; ++i)
-      SET_VECTOR_ELT(newx, i, xd[i]);
-  } break;
-  default : // # nocov
-    internal_error(__func__, "type '%s' not supported", type2char(TYPEOF(x)));  // # nocov
-  }
+  copyVectorElements(newx, x, (R_xlen_t)len, false, __func__);
   // if (verbose) Rprintf(_("Growing vector from %d to %d items of type '%s'\n"), len, newlen, type2char(TYPEOF(x)));
   // Would print for every column if here. Now just up in dogroups (one msg for each table grow).
   SHALLOW_DUPLICATE_ATTRIB(newx, x);

src/fifelse.c

@@ -96,7 +96,7 @@ SEXP fifelseR(SEXP l, SEXP a, SEXP b, SEXP na) {
   const int64_t bmask = len2>1 ? INT64_MAX : 0;
   const int64_t nmask = len3>1 ? INT64_MAX : 0;
 
-  const int *restrict pl = LOGICAL(l);
+  const int *restrict pl = LOGICAL_RO(l);
   SEXP ans = PROTECT(allocVector(tans, len0)); nprotect++;
   if (!na_a) 
     copyMostAttrib(a, ans);
@@ -237,7 +237,7 @@ SEXP fcaseR(SEXP rho, SEXP args) {
     if (!isLogical(whens)) {
       error(_("Argument #%d must be logical but was of type %s."), 2*i+1, type2char(TYPEOF(whens)));
     }
-    const int *restrict pwhens = LOGICAL(whens);
+    const int *restrict pwhens = LOGICAL_RO(whens);
     l = 0;
     if (i == 0) {
       n_ans = xlength(whens);
@@ -306,7 +306,7 @@ SEXP fcaseR(SEXP rho, SEXP args) {
     switch(TYPEOF(ans)) {
     case LGLSXP: {
       const int *restrict pthens;
-      if (!naout) pthens = LOGICAL(thens); // the content is not useful if out is NA_LOGICAL scalar
+      if (!naout) pthens = LOGICAL_RO(thens); // the content is not useful if out is NA_LOGICAL scalar
       int *restrict pans = LOGICAL(ans);
       const int pna = NA_LOGICAL;
       for (int64_t j=0; j<n_undecided; ++j) {
@@ -323,7 +323,7 @@ SEXP fcaseR(SEXP rho, SEXP args) {
     } break;
     case INTSXP: {
       const int *restrict pthens;
-      if (!naout) pthens = INTEGER(thens); // the content is not useful if out is NA_LOGICAL scalar
+      if (!naout) pthens = INTEGER_RO(thens); // the content is not useful if out is NA_LOGICAL scalar
       int *restrict pans = INTEGER(ans);
       const int pna = NA_INTEGER;
       for (int64_t j=0; j<n_undecided; ++j) {
@@ -340,7 +340,7 @@ SEXP fcaseR(SEXP rho, SEXP args) {
     } break;
     case REALSXP: {
       const double *restrict pthens;
-      if (!naout) pthens = REAL(thens); // the content is not useful if out is NA_LOGICAL scalar
+      if (!naout) pthens = REAL_RO(thens); // the content is not useful if out is NA_LOGICAL scalar
       double *restrict pans = REAL(ans);
       const double na_double = INHERITS(ans, char_integer64) ? NA_INT64_D : NA_REAL;
       const double pna = na_double;
@@ -358,7 +358,7 @@ SEXP fcaseR(SEXP rho, SEXP args) {
     } break;
     case CPLXSXP: {
       const Rcomplex *restrict pthens;
-      if (!naout) pthens = COMPLEX(thens); // the content is not useful if out is NA_LOGICAL scalar
+      if (!naout) pthens = COMPLEX_RO(thens); // the content is not useful if out is NA_LOGICAL scalar
       Rcomplex *restrict pans = COMPLEX(ans);
       const Rcomplex pna = NA_CPLX;
       for (int64_t j=0; j<n_undecided; ++j) {

src/fmelt.c

@@ -19,7 +19,7 @@ SEXP set_diff(SEXP x, int n) {
   if (n <= 0) error(_("'n' must be a positive integer"));
   SEXP table = PROTECT(seq_int(n, 1));       // TODO: using match to 1:n seems odd here, why use match at all
   SEXP xmatch = PROTECT(match(x, table, 0)); // Old comment:took a while to realise: matches vec against x - thanks to comment from Matt in assign.c!
-  const int *ixmatch = INTEGER(xmatch);
+  const int *ixmatch = INTEGER_RO(xmatch);
   int *buf = (int *) R_alloc(n, sizeof(*buf));
   int j=0;
   for (int i=0; i<n; ++i) {
@@ -39,7 +39,7 @@ SEXP which(SEXP x, Rboolean val) {
   int j=0, n = length(x);
   SEXP ans;
   if (!isLogical(x)) error(_("Argument to 'which' must be logical"));
-  const int *ix = LOGICAL(x);
+  const int *ix = LOGICAL_RO(x);
   int *buf = (int *) R_alloc(n, sizeof(*buf));
   for (int i=0; i<n; ++i) {
     if (ix[i] == val) {
@@ -70,7 +70,7 @@ static const char *concat(SEXP vec, SEXP idx) {
   int nidx=length(idx), nvec=length(vec);
   ans[0]='\0';
   if (nidx==0) return ans;
-  const int *iidx = INTEGER(idx);
+  const int *iidx = INTEGER_RO(idx);
   for (int i=0; i<nidx; ++i) {
     if (iidx[i]<1 || iidx[i]>nvec)
       internal_error(__func__, "'idx' must take values between 1 and length(vec); 1 <= idx <= %d", nvec); // # nocov
@@ -150,7 +150,7 @@ static SEXP unlist_(SEXP xint) {
   int *ians = INTEGER(ans), k=0;
   for (int i=0; i<n; ++i) {
     SEXP tmp = VECTOR_ELT(xint, i);
-    const int *itmp = INTEGER(tmp), n2=length(tmp);
+    const int *itmp = INTEGER_RO(tmp), n2=length(tmp);
     for (int j=0; j<n2; ++j)
       ians[k++] = itmp[j];
   }
@@ -522,7 +522,7 @@ SEXP getvaluecols(SEXP DT, SEXP dtnames, Rboolean valfactor, Rboolean verbose, s
         int thislen = 0;
         if (data->narm) {
           SEXP thisidx = VECTOR_ELT(data->not_NA_indices, j);
-          ithisidx = INTEGER(thisidx);
+          ithisidx = INTEGER_RO(thisidx);
           thislen = length(thisidx);
         }
         size_t size = RTYPE_SIZEOF(thiscol);
@@ -600,7 +600,7 @@ SEXP getvarcols(SEXP DT, SEXP dtnames, Rboolean varfactor, Rboolean verbose, str
     if (!varfactor) {
       SET_VECTOR_ELT(ansvars, 0, target=allocVector(STRSXP, data->totlen));
       if (!data->measure_is_list) {//one value column to output.
-        const int *thisvaluecols = INTEGER(VECTOR_ELT(data->valuecols, 0));
+        const int *thisvaluecols = INTEGER_RO(VECTOR_ELT(data->valuecols, 0));
         for (int j=0, ansloc=0; j<data->lmax; ++j) {
           const int thislen = data->narm ? length(VECTOR_ELT(data->not_NA_indices, j)) : data->nrow;
           SEXP str = STRING_ELT(dtnames, thisvaluecols[j]-1);
@@ -622,7 +622,7 @@ SEXP getvarcols(SEXP DT, SEXP dtnames, Rboolean varfactor, Rboolean verbose, str
         SEXP thisvaluecols = VECTOR_ELT(data->valuecols, 0);
         int len = length(thisvaluecols);
         levels = PROTECT(allocVector(STRSXP, len)); protecti++;
-        const int *vd = INTEGER(thisvaluecols);
+        const int *vd = INTEGER_RO(thisvaluecols);
         for (int j=0; j<len; ++j) SET_STRING_ELT(levels, j, STRING_ELT(dtnames, vd[j]-1));
         SEXP m = PROTECT(chmatch(levels, levels, 0)); protecti++;  // do we have any dups?
         int numRemove = 0;  // remove dups and any for which narm and all-NA
@@ -706,7 +706,7 @@ SEXP getidcols(SEXP DT, SEXP dtnames, Rboolean verbose, struct processData *data
       if (data->narm) {
         for (int j=0; j<data->lmax; ++j) {
           SEXP thisidx = VECTOR_ELT(data->not_NA_indices, j);
-          const int *ithisidx = INTEGER(thisidx);
+          const int *ithisidx = INTEGER_RO(thisidx);
           const int thislen = length(thisidx);
           for (int k=0; k<thislen; ++k)
             dtarget[counter + k] = dthiscol[ithisidx[k]-1];
@@ -725,7 +725,7 @@ SEXP getidcols(SEXP DT, SEXP dtnames, Rboolean verbose, struct processData *data
       if (data->narm) {
         for (int j=0; j<data->lmax; ++j) {
           SEXP thisidx = VECTOR_ELT(data->not_NA_indices, j);
-          const int *ithisidx = INTEGER(thisidx);
+          const int *ithisidx = INTEGER_RO(thisidx);
           const int thislen = length(thisidx);
           for (int k=0; k<thislen; ++k)
             itarget[counter + k] = ithiscol[ithisidx[k]-1];
@@ -740,7 +740,7 @@ SEXP getidcols(SEXP DT, SEXP dtnames, Rboolean verbose, struct processData *data
       if (data->narm) {
         for (int j=0; j<data->lmax; ++j) {
           SEXP thisidx = VECTOR_ELT(data->not_NA_indices, j);
-          const int *ithisidx = INTEGER(thisidx);
+          const int *ithisidx = INTEGER_RO(thisidx);
           const int thislen = length(thisidx);
           for (int k=0; k<thislen; ++k)
             SET_STRING_ELT(target, counter + k, STRING_ELT(thiscol, ithisidx[k]-1));
@@ -760,7 +760,7 @@ SEXP getidcols(SEXP DT, SEXP dtnames, Rboolean verbose, struct processData *data
       if (data->narm) {
         for (int j=0; j<data->lmax; ++j) {
           SEXP thisidx = VECTOR_ELT(data->not_NA_indices, j);
-          const int *ithisidx = INTEGER(thisidx);
+          const int *ithisidx = INTEGER_RO(thisidx);
           const int thislen = length(thisidx);
           for (int k=0; k<thislen; ++k)
             SET_VECTOR_ELT(target, counter + k, VECTOR_ELT(thiscol, ithisidx[k]-1));