Simplify nrow() getter for data.frame following code review

Author: eddelbuettel

ChangeLog

@@ -1,10 +1,14 @@
+2025-12-22  Dirk Eddelbuettel  <[email protected]>
+
+	* inst/include/Rcpp/DataFrame.h (nrow): Simplified per #1430 discussion
+	relying on compact sequence also removing use of ATTRIB
+
 2025-12-21  Dirk Eddelbuettel  <[email protected]>
 
-	* inst/include/Rcpp/DataFrame.h (nrow): Rewritten using R_mapAttrib()
-	avoiding ATTRIB() likely to be removed by R 4.6.0
-	* inst/include/Rcpp/proxy/AttributeProxy.h (attributeNames): Idem
-	* src/utilities.cpp: Add two helper functions used by R_mapAttrib()
-	* inst/include/Rcpp/routines.h: Register two new helper functions
+	* inst/include/Rcpp/proxy/AttributeProxy.h (attributeNames):
+	Rewritten using R_mapAttrib() avoiding ATTRIB() to be removed in R 4.6.0
+	* src/utilities.cpp: Add helper function used by R_mapAttrib()
+	* inst/include/Rcpp/routines.h: Register new helper function
 	* src/rcpp_init.cpp (registerFunctions): Idem
 
 	* .github/workflows/docker.yaml: Add workflow_dispatch, update

inst/include/Rcpp/DataFrame.h

@@ -61,36 +61,18 @@ namespace Rcpp{
         }
 
         // By definition, the number of rows in a data.frame is contained
-        // in its row.names attribute. If it has row names of the form 1:n,
-        // they will be stored as {NA_INTEGER, -<nrow>}. Unfortunately,
-        // getAttrib(df, R_RowNamesSymbol) will force an expansion of that
-        // compact form thereby allocating a huge vector when we just want
-        // the row.names. Hence this workaround.
-        inline int nrow() const {
-#if R_VERSION >= R_Version(4, 6, 0)
-            Shield<SEXP> v = R_mapAttrib(Parent::get__(), get_row_count, R_NilValue);
-            if (v != NULL && TYPEOF(v) == INTSXP) {
-                return INTEGER(v)[0];
-            } else {
-                // TODO: error?
-                return NA_INTEGER;
-            }
-#else
-            SEXP rn = R_NilValue ;
-            SEXP att = ATTRIB( Parent::get__() )  ;
-            while( att != R_NilValue ){
-                if( TAG(att) == R_RowNamesSymbol ) {
-                    rn = CAR(att) ;
-                    break ;
-                }
-                att = CDR(att) ;
-            }
-            if (Rf_isNull(rn))
-                return 0;
-            if (TYPEOF(rn) == INTSXP && LENGTH(rn) == 2 && INTEGER(rn)[0] == NA_INTEGER)
-                return std::abs(INTEGER(rn)[1]);
-            return LENGTH(rn);
-#endif
+        // in its row.names attribute. Since R 3.5.0 this is returned as a
+        // compact sequence from which we can just take the length
+        //   Shield<SEXP> rn = Rf_getAttrib(Parent::get__(), R_RowNamesSymbol);
+        //   return Rf_xlength(rn);
+        // But as this makes an allocation an even simpler check on length as
+        // discussed in #1430 is also possible and preferable. We also switch
+        // to returning R_xlen_t which as upcast from int is safe
+        inline R_xlen_t nrow() const {
+            SEXP x = Parent::get__();
+            // We can simplify: zero row DFs have no names, else length of all vector same
+            // by design constraint so can use vector length for desired row count
+            return (XLENGTH(x) == 0) ? 0 : XLENGTH(VECTOR_ELT(x, 0));
         }
 
         template <typename T>
@@ -118,8 +100,8 @@ namespace Rcpp{
         }
 
         // Offer multiple variants to accomodate both old interface here and signatures in other classes
-        inline int nrows() const { return DataFrame_Impl::nrow(); }
-        inline int rows()  const { return DataFrame_Impl::nrow(); }
+        inline R_xlen_t nrows() const { return DataFrame_Impl::nrow(); }
+        inline R_xlen_t rows()  const { return DataFrame_Impl::nrow(); }
 
         inline R_xlen_t ncol()  const { return DataFrame_Impl::length(); }
         inline R_xlen_t cols()  const { return DataFrame_Impl::length(); }

inst/include/Rcpp/routines.h

@@ -317,13 +317,6 @@ inline attribute_hidden SEXP get_attr_names(SEXP tag, SEXP attr, void* data){
     return fun(tag, attr, data);
 }
 
-inline attribute_hidden SEXP get_row_count(SEXP tag, SEXP attr, void* data){
-    typedef SEXP (*Fun)(SEXP, SEXP, void*);
-    static Fun fun = GET_CALLABLE("get_row_count");
-    return fun(tag, attr, data);
-}
-
-
 #endif
 
 

src/rcpp_init.cpp

@@ -129,7 +129,6 @@ void registerFunctions(){
     RCPP_REGISTER(Rcpp_cerr_get)
 
     RCPP_REGISTER(get_attr_names)
-    RCPP_REGISTER(get_row_count)
     #undef RCPP_REGISTER
 }
 

src/utilities.cpp

@@ -34,22 +34,3 @@ SEXP get_attr_names(SEXP tag, SEXP attr, void* data) {
     vecptr->push_back(s);
     return NULL;
 }
-
-// See WRE Section 6.22.6 'Working with attributes' under R 4.6.0 or later
-// This function extracts the number of rows in a data frame
-// It is used DataFrame_impl::nrow() in a call to R_mapAttrib()
-// [[Rcpp::register]]
-SEXP get_row_count(SEXP tag, SEXP attr, void* data) {
-    if (tag == R_RowNamesSymbol) {
-        if (TYPEOF(attr) == INTSXP && LENGTH(attr) == 2 && INTEGER(attr)[0] == NA_INTEGER) {
-            int n = std::abs(INTEGER(attr)[1]);
-            //Rcpp::Rcout << "Seeing " << n << std::endl;
-            return Rf_ScalarInteger(n);
-        }
-        if (Rf_isNull(attr)) {
-            return Rf_ScalarInteger(0);
-        }
-        return Rf_ScalarInteger(LENGTH(attr));
-    }
-    return NULL;
-}