From 537ad7e6dae1d7f370a6d876a753260c578574c5 Mon Sep 17 00:00:00 2001
From: Sebastian Krantz <sebastian.krantz@graduateinstitute.ch>
Date: Sun, 23 Jun 2024 14:49:46 +0200
Subject: [PATCH] Replace STRING_PTR.

---
 src/TRA.c                  | 38 +++++++++++++++++++-------------------
 src/base_radixsort.c       |  2 +-
 src/base_radixsort.h       |  2 ++
 src/data.table.h           |  5 ++---
 src/data.table_rbindlist.c | 16 ++++++++--------
 src/data.table_subset.c    | 10 +++++-----
 src/data.table_utils.c     | 10 +++++-----
 src/ffirst.c               | 12 ++++++------
 src/flast.c                | 14 +++++++-------
 src/fmode.c                | 18 +++++++++---------
 src/fndistinct.c           | 12 ++++++------
 src/fnobs.c                |  8 ++++----
 src/gsplit.c               | 16 ++++++++--------
 src/join.c                 |  4 ++--
 src/kit.h                  |  4 +++-
 src/kit_dup.c              | 16 ++++++++--------
 src/match.c                | 22 +++++++++++-----------
 src/pivot.c                |  6 +++---
 src/programming.c          | 24 ++++++++++++------------
 src/small_helper.c         |  8 ++++----
 20 files changed, 125 insertions(+), 122 deletions(-)

diff --git a/src/TRA.c b/src/TRA.c
index 154c8ca6..92b3bf09 100644
--- a/src/TRA.c
+++ b/src/TRA.c
@@ -125,13 +125,13 @@ SEXP ret1(SEXP x, SEXP xAG, SEXP g, int set) {
     }
     case STRSXP:
     {
-      SEXP *pout = STRING_PTR(out);
+      SEXP *pout = SEXPPTR(out);
       if(nog) {
         SEXP AG = asChar(xAG);
         #pragma omp simd
         for(int i = 0; i < l; ++i) pout[i] = AG;
       } else {
-        SEXP *AG = STRING_PTR(xAG)-1;
+        SEXP *AG = SEXPPTR(xAG)-1;
         #pragma omp simd
         for(int i = 0; i < l; ++i) pout[i] = AG[pg[i]];
       }
@@ -237,13 +237,13 @@ SEXP ret2(SEXP x, SEXP xAG, SEXP g, int set) {
         break;
       }
       case STRSXP: {
-        SEXP *pout = STRING_PTR(out);
+        SEXP *pout = SEXPPTR(out);
         if(nog) {
           SEXP AG = asChar(xAG);
           #pragma omp simd
           for(int i = 0; i < l; ++i) pout[i] = ISNAN(px[i]) ? NA_STRING : AG;
         } else {
-          SEXP *AG = STRING_PTR(xAG)-1;
+          SEXP *AG = SEXPPTR(xAG)-1;
           #pragma omp simd
           for(int i = 0; i < l; ++i) pout[i] = ISNAN(px[i]) ? NA_STRING : AG[pg[i]];
         }
@@ -287,13 +287,13 @@ SEXP ret2(SEXP x, SEXP xAG, SEXP g, int set) {
         break;
       }
       case STRSXP: {
-        SEXP *pout = STRING_PTR(out);
+        SEXP *pout = SEXPPTR(out);
         if(nog) {
           SEXP AG = asChar(xAG);
           #pragma omp simd
           for(int i = 0; i < l; ++i) pout[i] = (px[i] == NA_INTEGER) ? NA_STRING : AG;
         } else {
-          SEXP *AG = STRING_PTR(xAG)-1;
+          SEXP *AG = SEXPPTR(xAG)-1;
           #pragma omp simd
           for(int i = 0; i < l; ++i) pout[i] = (px[i] == NA_INTEGER) ? NA_STRING : AG[pg[i]];
         }
@@ -306,7 +306,7 @@ SEXP ret2(SEXP x, SEXP xAG, SEXP g, int set) {
   }
   case STRSXP:
   {
-    SEXP *px = STRING_PTR(x);
+    SEXP *px = SEXPPTR(x);
     switch(txAG) {
       case REALSXP: {
         double *pout = REAL(out);
@@ -336,13 +336,13 @@ SEXP ret2(SEXP x, SEXP xAG, SEXP g, int set) {
         break;
       }
       case STRSXP: {
-        SEXP *pout = STRING_PTR(out);
+        SEXP *pout = SEXPPTR(out);
         if(nog) {
           SEXP AG = asChar(xAG);
           #pragma omp simd
           for(int i = 0; i < l; ++i) pout[i] = (px[i] == NA_STRING) ? NA_STRING : AG;
         } else {
-          SEXP *AG = STRING_PTR(xAG)-1;
+          SEXP *AG = SEXPPTR(xAG)-1;
           #pragma omp simd
           for(int i = 0; i < l; ++i) pout[i] = (px[i] == NA_STRING) ? NA_STRING : AG[pg[i]];
         }
@@ -448,7 +448,7 @@ SEXP ret0(SEXP x, SEXP xAG, SEXP g, int set) {
     }
     case STRSXP:
     {
-      SEXP *px = STRING_PTR(x), *pout = STRING_PTR(out);
+      SEXP *px = SEXPPTR(x), *pout = SEXPPTR(out);
       if(nog) {
         SEXP AG = asChar(xAG);
         #pragma omp simd
@@ -459,7 +459,7 @@ SEXP ret0(SEXP x, SEXP xAG, SEXP g, int set) {
           case LGLSXP:
           case INTSXP: error("Cannot replace missing values in string with numeric data");
           case STRSXP: {
-            SEXP *AG = STRING_PTR(xAG)-1;
+            SEXP *AG = SEXPPTR(xAG)-1;
             #pragma omp simd
             for(int i = 0; i < l; ++i) pout[i] = (px[i] == NA_STRING) ? AG[pg[i]] : px[i];
             break;
@@ -728,7 +728,7 @@ SEXP TRAlC(SEXP x, SEXP xAG, SEXP g, SEXP Rret, SEXP Rset) {
       break;
     }
     case STRSXP: {
-      SEXP *pAG = STRING_PTR(xAG);
+      SEXP *pAG = SEXPPTR(xAG);
       RETLOOPS(ScalarString(pAG[j]))
       break;
     }
@@ -814,7 +814,7 @@ SEXP TRAmC(SEXP x, SEXP xAG, SEXP g, SEXP Rret, SEXP Rset) {
           }
           case STRSXP:
           {
-            SEXP *pout = STRING_PTR(out), *pAG = STRING_PTR(xAG);
+            SEXP *pout = SEXPPTR(out), *pAG = SEXPPTR(xAG);
             if(nog) {
               for(int j = 0; j != col; ++j) {
                 int s = j * row, e = s + row;
@@ -883,7 +883,7 @@ SEXP TRAmC(SEXP x, SEXP xAG, SEXP g, SEXP Rret, SEXP Rset) {
               }
               case STRSXP:
               {
-                SEXP *pout = STRING_PTR(out), *pAG = STRING_PTR(xAG);
+                SEXP *pout = SEXPPTR(out), *pAG = SEXPPTR(xAG);
                 if(nog) {
                   for(int j = 0; j != col; ++j) {
                     int s = j * row, e = s + row;
@@ -951,7 +951,7 @@ SEXP TRAmC(SEXP x, SEXP xAG, SEXP g, SEXP Rret, SEXP Rset) {
               }
               case STRSXP:
               {
-                SEXP *pout = STRING_PTR(out), *pAG = STRING_PTR(xAG);
+                SEXP *pout = SEXPPTR(out), *pAG = SEXPPTR(xAG);
                 if(nog) {
                   for(int j = 0; j != col; ++j) {
                     int s = j * row, e = s + row;
@@ -975,7 +975,7 @@ SEXP TRAmC(SEXP x, SEXP xAG, SEXP g, SEXP Rret, SEXP Rset) {
           }
           case STRSXP:
           {
-            SEXP *px = STRING_PTR(x);
+            SEXP *px = SEXPPTR(x);
             switch(txAG) {
               case REALSXP:
               {
@@ -1018,7 +1018,7 @@ SEXP TRAmC(SEXP x, SEXP xAG, SEXP g, SEXP Rret, SEXP Rset) {
               }
               case STRSXP:
               {
-                SEXP *pout = STRING_PTR(out), *pAG = STRING_PTR(xAG);
+                SEXP *pout = SEXPPTR(out), *pAG = SEXPPTR(xAG);
                 if(nog) {
                   for(int j = 0; j != col; ++j) {
                     int s = j * row, e = s + row;
@@ -1165,14 +1165,14 @@ SEXP TRAmC(SEXP x, SEXP xAG, SEXP g, SEXP Rret, SEXP Rset) {
       }
       case STRSXP:
       {
-        SEXP *px = STRING_PTR(x), *pout = STRING_PTR(out);
+        SEXP *px = SEXPPTR(x), *pout = SEXPPTR(out);
         switch(txAG) {
         case REALSXP:
         case INTSXP:
         case LGLSXP: error("Cannot replace missing values in string with numeric data");
         case STRSXP:
         {
-          SEXP *pAG = STRING_PTR(xAG);
+          SEXP *pAG = SEXPPTR(xAG);
           if(nog) {
             for(int j = 0; j != col; ++j) {
               int s = j * row, e = s + row;
diff --git a/src/base_radixsort.c b/src/base_radixsort.c
index 3ef55f1c..c217d153 100644
--- a/src/base_radixsort.c
+++ b/src/base_radixsort.c
@@ -931,7 +931,7 @@ static int StrCmp(SEXP x, SEXP y)            // also used by bmerge and chmatch
 void checkEncodings(SEXP x) // static
 {
   cetype_t ce;
-  SEXP *px = STRING_PTR(x);
+  SEXP *px = SEXPPTR(x);
   int i, lx = length(x);
   for (i = 0; i != lx && px[i] == NA_STRING; ++i);
 
diff --git a/src/base_radixsort.h b/src/base_radixsort.h
index aa4858b8..2d5283f3 100644
--- a/src/base_radixsort.h
+++ b/src/base_radixsort.h
@@ -8,6 +8,8 @@
 // #define IS_ASCII(x) ((x)->sxpinfo.gp & ASCII_MASK)
 // #define IS_ASCII(x) (LEVELS(x) & ASCII_MASK)
 
+#define SEXPPTR(x) ((SEXP *)DATAPTR(x))  // Replacing STRING_PTR
+
 // NOTE: All of this is copied from Defn.h: https://github.com/wch/r-source/blob/28de75af0541f93832c5899139b969d290bf422e/src/include/Defn.h
 // to avoid checking for ALTREP in TRUELENGTH, which slows down the code unnecessarily...
 
diff --git a/src/data.table.h b/src/data.table.h
index 42be1773..3a1e348a 100644
--- a/src/data.table.h
+++ b/src/data.table.h
@@ -3,10 +3,9 @@
  and licensed under a Mozilla Public License 2.0 (MPL-2.0) license.
 */
 
-#include <R.h>
 #define USE_RINTERNALS
-#include <Rinternals.h>
-#include <stdint.h> // for uint64_t rather than unsigned long long
+#include "base_radixsort.h"
+// #include <stdint.h> // for uint64_t rather than unsigned long long
 #include <stdbool.h>
 // #include "types.h"
 
diff --git a/src/data.table_rbindlist.c b/src/data.table_rbindlist.c
index 05018ea1..09e3829b 100644
--- a/src/data.table_rbindlist.c
+++ b/src/data.table_rbindlist.c
@@ -38,7 +38,7 @@ void writeNA(SEXP v, const int from, const int n)
     for (int i=from; i<=to; ++i) vd[i] = NA_CPLX;
   } break;
   case STRSXP: {
-    SEXP *vd = STRING_PTR(v);
+    SEXP *vd = SEXPPTR(v);
     for (int i=from; i<=to; ++i) vd[i] = NA_STRING;
   } break;
   case VECSXP:
@@ -262,7 +262,7 @@ SEXP rbindlist(SEXP l, SEXP usenamesArg, SEXP fillArg, SEXP idcolArg)
       if (isNull(li) || !LENGTH(li)) continue;
       const SEXP cn = getAttrib(li, R_NamesSymbol);
       if (!length(cn)) continue;
-      const SEXP *cnp = STRING_PTR(cn);
+      const SEXP *cnp = SEXPPTR(cn);
       for (int j=0; j<thisncol; j++) {
         SEXP s = cnp[j];
         if (TRUELENGTH(s)<0) continue;  // seen this name before
@@ -293,7 +293,7 @@ SEXP rbindlist(SEXP l, SEXP usenamesArg, SEXP fillArg, SEXP idcolArg)
       if (thisncol==0) continue;
       const SEXP cn = getAttrib(li, R_NamesSymbol);
       if (!length(cn)) continue;
-      const SEXP *cnp = STRING_PTR(cn);
+      const SEXP *cnp = SEXPPTR(cn);
       memset(counts, 0, nuniq*sizeof(int));
       for (int j=0; j<thisncol; j++) {
         SEXP s = cnp[j];
@@ -333,7 +333,7 @@ SEXP rbindlist(SEXP l, SEXP usenamesArg, SEXP fillArg, SEXP idcolArg)
       if (!length(cn)) {
         for (int j=0; j<thisncol; j++) colMapRaw[i*ncol + j] = j;
       } else {
-        const SEXP *cnp = STRING_PTR(cn);
+        const SEXP *cnp = SEXPPTR(cn);
         memset(counts, 0, nuniq*sizeof(int));
         for (int j=0; j<thisncol; j++) {
           SEXP s = cnp[j];
@@ -540,7 +540,7 @@ SEXP rbindlist(SEXP l, SEXP usenamesArg, SEXP fillArg, SEXP idcolArg)
         //      c( a<c<b, c<b, 'c,b'   ) => a<c<b  because the regular factor/character items c and b exist in the ordered levels
         //      c( a<c<b, c<b, 'c,d'   ) => a<c<b<d  'd' from non-ordered item added on the end of longest ordered levels
         //      c( a<c<b, c<b<d<e )  => regular factor because this case isn't yet implemented. a<c<b<d<e would be possible in future (extending longest at the beginning or end)
-        const SEXP *sd = STRING_PTR(longestLevels);
+        const SEXP *sd = SEXPPTR(longestLevels);
         nLevel = allocLevel = longestLen;
         levelsRaw = (SEXP *)malloc(nLevel * sizeof(SEXP));
         if (!levelsRaw) { savetl_end(); error("Failed to allocate working memory for %d ordered factor levels of result column %d", nLevel, idcol+j+1); }
@@ -556,7 +556,7 @@ SEXP rbindlist(SEXP l, SEXP usenamesArg, SEXP fillArg, SEXP idcolArg)
           SEXP thisCol = VECTOR_ELT(VECTOR_ELT(l, i), w);
           if (isOrdered(thisCol)) {
             SEXP levels = getAttrib(thisCol, R_LevelsSymbol);
-            const SEXP *levelsD = STRING_PTR(levels);
+            const SEXP *levelsD = SEXPPTR(levels);
             const int n = length(levels);
             for (int k=0, last=0; k<n; ++k) {
               SEXP s = levelsD[k];
@@ -595,7 +595,7 @@ SEXP rbindlist(SEXP l, SEXP usenamesArg, SEXP fillArg, SEXP idcolArg)
           SEXP thisCol = VECTOR_ELT(li, w);
           SEXP thisColStr = isFactor(thisCol) ? getAttrib(thisCol, R_LevelsSymbol) : (isString(thisCol) ? thisCol : VECTOR_ELT(coercedForFactor, i));
           const int n = length(thisColStr);
-          const SEXP *thisColStrD = STRING_PTR(thisColStr);  // D for data
+          const SEXP *thisColStrD = SEXPPTR(thisColStr);  // D for data
           for (int k=0; k<n; ++k) {
             SEXP s = thisColStrD[k];
             if (s==NA_STRING ||             // remove NA from levels; test 1979 found by package emil when revdep testing 1.12.2 (#3473)
@@ -662,7 +662,7 @@ SEXP rbindlist(SEXP l, SEXP usenamesArg, SEXP fillArg, SEXP idcolArg)
               }
             }
           } else {
-            const SEXP *sd = STRING_PTR(thisColStr);
+            const SEXP *sd = SEXPPTR(thisColStr);
             if (length(thisCol)<=1) {
               const int val = (length(thisCol)==1 && sd[0]!=NA_STRING) ? -TRUELENGTH(sd[0]) : NA_INTEGER;
               for (int r=0; r<thisnrow; ++r) targetd[ansloc+r] = val;
diff --git a/src/data.table_subset.c b/src/data.table_subset.c
index a66a1d7b..45d5e6fa 100644
--- a/src/data.table_subset.c
+++ b/src/data.table_subset.c
@@ -102,8 +102,8 @@ static SEXP shallow(SEXP dt, SEXP cols, R_len_t n)
   SEXP names = PROTECT(getAttrib(dt, R_NamesSymbol)); protecti++;
   SEXP newnames = PROTECT(allocVector(STRSXP, n)); protecti++;
 
-  const SEXP *pdt = SEXPPTR_RO(dt), *pnam = STRING_PTR(names);
-  SEXP *pnewdt = SEXPPTR(newdt), *pnnam = STRING_PTR(newnames);
+  const SEXP *pdt = SEXPPTR_RO(dt), *pnam = SEXPPTR(names);
+  SEXP *pnewdt = SEXPPTR(newdt), *pnnam = SEXPPTR(newnames);
 
   const int l = isNull(cols) ? LENGTH(dt) : length(cols);
   if (isNull(cols)) {
@@ -215,7 +215,7 @@ void subsetVectorRaw(SEXP ans, SEXP source, SEXP idx, const bool anyNA)
       // TODO - discuss with Luke Tierney. Produce benchmarks on integer/double to see if it's worth making a safe
     //        API interface for package use for STRSXP.
     // Aside: setkey() is a separate special case (a permutation) and does do this in parallel without using SET_*.
-    SEXP *restrict sp = STRING_PTR(source)-1, *restrict ap = STRING_PTR(ans);
+    SEXP *restrict sp = SEXPPTR(source)-1, *restrict ap = SEXPPTR(ans);
     PARLOOP(NA_STRING);
   } break;
   case VECSXP : {
@@ -423,7 +423,7 @@ SEXP subsetCols(SEXP x, SEXP cols, SEXP checksf) { // SEXP fretall
   // sf data frames: Need to add sf_column
   if(oxl && asLogical(checksf) && INHERITS(x, char_sf)) {
     int sfcoln = NA_INTEGER, sf_col_sel = 0;
-    SEXP *pnam = STRING_PTR(nam), sfcol = asChar(getAttrib(x, sym_sf_column));
+    SEXP *pnam = SEXPPTR(nam), sfcol = asChar(getAttrib(x, sym_sf_column));
     for(int i = l; i--; ) {
       if(pnam[i] == sfcol) {
         sfcoln = i + 1;
@@ -511,7 +511,7 @@ SEXP subsetDT(SEXP x, SEXP rows, SEXP cols, SEXP checkrows) { // , SEXP fastret
       if(oxl && INHERITS(x, char_sf)) {
         int sfcoln = NA_INTEGER, sf_col_sel = 0;
         SEXP nam = PROTECT(getAttrib(x, R_NamesSymbol));
-        SEXP *pnam = STRING_PTR(nam), sfcol = asChar(getAttrib(x, sym_sf_column));
+        SEXP *pnam = SEXPPTR(nam), sfcol = asChar(getAttrib(x, sym_sf_column));
         for(int i = l; i--; ) {
           if(pnam[i] == sfcol) {
             sfcoln = i + 1;
diff --git a/src/data.table_utils.c b/src/data.table_utils.c
index 3be12f8b..9d49c7df 100644
--- a/src/data.table_utils.c
+++ b/src/data.table_utils.c
@@ -16,7 +16,7 @@ SEXP setnames(SEXP x, SEXP nam) {
       return x;
     }
     SEXP newnam = PROTECT(allocVector(STRSXP, n)),
-      *pnn = STRING_PTR(newnam), *pn = STRING_PTR(nam);
+      *pnn = SEXPPTR(newnam), *pn = SEXPPTR(nam);
     for(int i = 0; i < l; ++i) pnn[i] = pn[i];
     SETLENGTH(newnam, l);
     SET_TRUELENGTH(newnam, n);
@@ -58,7 +58,7 @@ bool allNA(SEXP x, bool errorForBadType) {
     }
     return true;
   case STRSXP: {
-    const SEXP *xd = STRING_PTR(x);
+    const SEXP *xd = SEXPPTR(x);
     for (int i=0; i != n; ++i)    if (xd[i]!=NA_STRING) {
       return false;
     }
@@ -159,7 +159,7 @@ SEXP dt_na(SEXP x, SEXP cols, SEXP Rprop, SEXP Rcount) {
         for (int j=0; j != n; ++j) ians[j] += (iv[j] == NA_INTEGER);
       } break;
       case STRSXP: {
-        const SEXP *sv = STRING_PTR(v);
+        const SEXP *sv = SEXPPTR(v);
         for (int j=0; j != n; ++j) ians[j] += (sv[j] == NA_STRING);
       } break;
       case REALSXP: {
@@ -204,7 +204,7 @@ SEXP dt_na(SEXP x, SEXP cols, SEXP Rprop, SEXP Rcount) {
         for (int j=0; j != n; ++j) ians[j] |= (iv[j] == NA_INTEGER);
       } break;
       case STRSXP: {
-        const SEXP *sv = STRING_PTR(v);
+        const SEXP *sv = SEXPPTR(v);
         for (int j=0; j != n; ++j) ians[j] |= (sv[j] == NA_STRING);
       } break;
       case REALSXP: {
@@ -294,7 +294,7 @@ SEXP setcolorder(SEXP x, SEXP o) {
   }
   Free(seen);
 
-  SEXP *tmp = Calloc(ncol, SEXP), *namesd = STRING_PTR(names);
+  SEXP *tmp = Calloc(ncol, SEXP), *namesd = SEXPPTR(names);
   const SEXP *xd = SEXPPTR_RO(x);
   for (int i=0; i != ncol; ++i) tmp[i] = xd[od[i]-1];
   for (int i=0; i != ncol; ++i) SET_VECTOR_ELT(x, i, tmp[i]);
diff --git a/src/ffirst.c b/src/ffirst.c
index 9fa52dbf..33a5e493 100644
--- a/src/ffirst.c
+++ b/src/ffirst.c
@@ -25,7 +25,7 @@ SEXP ffirst_impl(SEXP x, int ng, SEXP g, int narm, int *gl) {
           break;
         }
         case STRSXP: {
-          SEXP *px = STRING_PTR(x);
+          SEXP *px = SEXPPTR(x);
           while(px[j] == NA_STRING && j != end) ++j;
           SET_STRING_ELT(out, 0, px[j]);
           break;
@@ -86,7 +86,7 @@ SEXP ffirst_impl(SEXP x, int ng, SEXP g, int narm, int *gl) {
         break;
       }
       case STRSXP: {
-        SEXP *px = STRING_PTR(x), *pout = STRING_PTR(out);
+        SEXP *px = SEXPPTR(x), *pout = SEXPPTR(out);
         for(int i = ng; i--; ) pout[i] = NA_STRING;
         --pout;
         for(int i = 0; i != l; ++i) {
@@ -156,7 +156,7 @@ SEXP ffirst_impl(SEXP x, int ng, SEXP g, int narm, int *gl) {
         break;
       }
       case STRSXP:{
-        SEXP *px = STRING_PTR(x)-1, *pout = STRING_PTR(out);
+        SEXP *px = SEXPPTR(x)-1, *pout = SEXPPTR(out);
         for(int i = ng; i--; ) pout[i] = gl[i] == NA_INTEGER ? NA_STRING : px[gl[i]];
         break;
       }
@@ -255,7 +255,7 @@ SEXP ffirstmC(SEXP x, SEXP Rng, SEXP g, SEXP gst, SEXP Rnarm, SEXP Rdrop) {
         break;
       }
       case STRSXP: {
-        SEXP *px = STRING_PTR(x), *pout = STRING_PTR(out);
+        SEXP *px = SEXPPTR(x), *pout = SEXPPTR(out);
         for(int j = 0, i = 0; j != col; ++j) {
           while(px[i] == NA_STRING && i != end) ++i;
           pout[j] = px[i]; px += l; i = 0;
@@ -326,7 +326,7 @@ SEXP ffirstmC(SEXP x, SEXP Rng, SEXP g, SEXP gst, SEXP Rnarm, SEXP Rdrop) {
         break;
       }
       case STRSXP: {
-        SEXP *px = STRING_PTR(x), *pout = STRING_PTR(out);
+        SEXP *px = SEXPPTR(x), *pout = SEXPPTR(out);
         for(int i = ng * col; i--; ) pout[i] = NA_STRING;
         --pout;
         for(int j = 0; j != col; ++j) {
@@ -389,7 +389,7 @@ SEXP ffirstmC(SEXP x, SEXP Rng, SEXP g, SEXP gst, SEXP Rnarm, SEXP Rdrop) {
         break;
       }
       case STRSXP: {
-        SEXP *px = STRING_PTR(x)-1, *pout = STRING_PTR(out);
+        SEXP *px = SEXPPTR(x)-1, *pout = SEXPPTR(out);
         for(int j = 0; j != col; ++j) {
           for(int i = ng; i--; ) pout[i] = pgl[i] == NA_INTEGER ? NA_STRING : px[pgl[i]];
           px += l; pout += ng;
diff --git a/src/flast.c b/src/flast.c
index 831ce399..9b99ef91 100644
--- a/src/flast.c
+++ b/src/flast.c
@@ -17,7 +17,7 @@ SEXP flast_impl(SEXP x, int ng, SEXP g, int narm, int *gl) {
           break;
         }
         case STRSXP: {
-          SEXP *px = STRING_PTR(x);
+          SEXP *px = SEXPPTR(x);
           while(px[j] == NA_STRING && j != 0) --j;
           SET_STRING_ELT(out, 0, px[j]);
           break;
@@ -78,7 +78,7 @@ SEXP flast_impl(SEXP x, int ng, SEXP g, int narm, int *gl) {
         break;
       }
       case STRSXP: {
-        SEXP *px = STRING_PTR(x), *pout = STRING_PTR(out);
+        SEXP *px = SEXPPTR(x), *pout = SEXPPTR(out);
         for(int i = ng; i--; ) pout[i] = NA_STRING;
         --pout;
         for(int i = l; i--; ) {
@@ -137,7 +137,7 @@ SEXP flast_impl(SEXP x, int ng, SEXP g, int narm, int *gl) {
         break;
       }
       case STRSXP: {
-        SEXP *px = STRING_PTR(x), *pout = STRING_PTR(out);
+        SEXP *px = SEXPPTR(x), *pout = SEXPPTR(out);
         for(int i = ng; i--; ) pout[i] = gl[i] == NA_INTEGER ? NA_STRING : px[gl[i]];
         break;
       }
@@ -214,7 +214,7 @@ SEXP flastmC(SEXP x, SEXP Rng, SEXP g, SEXP Rnarm, SEXP Rdrop) {
         break;
       }
       case STRSXP: {
-        SEXP *px = STRING_PTR(x), *pout = STRING_PTR(out);
+        SEXP *px = SEXPPTR(x), *pout = SEXPPTR(out);
         for(int j = 0, i = l-1; j != col; ++j) {
           while(px[i] == NA_STRING && i != 0) --i;
           pout[j] = px[i]; px += l; i = l-1;
@@ -231,7 +231,7 @@ SEXP flastmC(SEXP x, SEXP Rng, SEXP g, SEXP Rnarm, SEXP Rdrop) {
         break;
       }
       case VECSXP: {
-        SEXP *px = STRING_PTR(x), *pout = STRING_PTR(out);
+        SEXP *px = SEXPPTR(x), *pout = SEXPPTR(out);
         for(int j = 0, i = l-1; j != col; ++j) {
           while(length(px[i]) == 0 && i != 0) --i;
           pout[j] = px[i]; px += l; i = l-1;
@@ -283,7 +283,7 @@ SEXP flastmC(SEXP x, SEXP Rng, SEXP g, SEXP Rnarm, SEXP Rdrop) {
         break;
       }
       case STRSXP: {
-        SEXP *px = STRING_PTR(x), *pout = STRING_PTR(out);
+        SEXP *px = SEXPPTR(x), *pout = SEXPPTR(out);
         for(int i = ng * col; i--; ) pout[i] = NA_STRING;
         --pout;
         for(int j = 0; j != col; ++j) {
@@ -342,7 +342,7 @@ SEXP flastmC(SEXP x, SEXP Rng, SEXP g, SEXP Rnarm, SEXP Rdrop) {
         break;
       }
       case STRSXP: {
-        SEXP *px = STRING_PTR(x), *pout = STRING_PTR(out);
+        SEXP *px = SEXPPTR(x), *pout = SEXPPTR(out);
         for(int j = 0; j != col; ++j) {
           for(int i = ng; i--; ) pout[i] = pgl[i] == NA_INTEGER ? NA_STRING : px[pgl[i]];
           px += l; pout += ng;
diff --git a/src/fmode.c b/src/fmode.c
index a67409e2..0f2bba87 100644
--- a/src/fmode.c
+++ b/src/fmode.c
@@ -637,7 +637,7 @@ SEXP mode_impl_plain(SEXP x, int narm, int ret) {
     case INTSXP:  return ScalarInteger(isFactor(x) ? mode_fct_logi(INTEGER(x), &l, l, nlevels(x), 1, narm, ret) :
                                     mode_int(INTEGER(x), &l, l, 1, narm, ret));
     case LGLSXP: return my_ScalarLogical(mode_fct_logi(LOGICAL(x), &l, l, 1, 1, narm, ret));
-    case STRSXP: return ScalarString(mode_string(STRING_PTR(x), &l, l, 1, narm, ret));
+    case STRSXP: return ScalarString(mode_string(SEXPPTR(x), &l, l, 1, narm, ret));
     default: error("Not Supported SEXP Type: '%s'", type2char(TYPEOF(x)));
   }
 }
@@ -659,7 +659,7 @@ SEXP w_mode_impl_plain(SEXP x, double *pw, int narm, int ret) {
     case INTSXP:  return ScalarInteger(isFactor(x) ? w_mode_fct_logi(INTEGER(x), pw, &l, l, nlevels(x), 1, narm, ret) :
                                     w_mode_int(INTEGER(x), pw, &l, l, 1, narm, ret));
     case LGLSXP:  return my_ScalarLogical(w_mode_fct_logi(LOGICAL(x), pw, &l, l, 1, 1, narm, ret));
-    case STRSXP:  return ScalarString(w_mode_string(STRING_PTR(x), pw, &l, l, 1, narm, ret));
+    case STRSXP:  return ScalarString(w_mode_string(SEXPPTR(x), pw, &l, l, 1, narm, ret));
     default: error("Not Supported SEXP Type: '%s'", type2char(TYPEOF(x)));
   }
 }
@@ -711,7 +711,7 @@ SEXP mode_g_impl(SEXP x, int ng, int *pgs, int *po, int *pst, int sorted, int na
         break;
       }
       case STRSXP: {
-        SEXP *px = STRING_PTR(x), *pres = STRING_PTR(res);
+        SEXP *px = SEXPPTR(x), *pres = SEXPPTR(res);
         #pragma omp parallel for num_threads(nthreads)
         for(int gr = 0; gr < ng; ++gr)
           pres[gr] = pgs[gr] == 0 ? NA_STRING : mode_string(px + pst[gr]-1, po, pgs[gr], 1, narm, ret);
@@ -750,7 +750,7 @@ SEXP mode_g_impl(SEXP x, int ng, int *pgs, int *po, int *pst, int sorted, int na
         break;
       }
       case STRSXP: {
-        SEXP *px = STRING_PTR(x), *pres = STRING_PTR(res);
+        SEXP *px = SEXPPTR(x), *pres = SEXPPTR(res);
         #pragma omp parallel for num_threads(nthreads)
         for(int gr = 0; gr < ng; ++gr)
           pres[gr] = pgs[gr] == 0 ? NA_STRING : mode_string(px, po + pst[gr]-1, pgs[gr], 0, narm, ret);
@@ -804,7 +804,7 @@ SEXP w_mode_g_impl(SEXP x, double *pw, int ng, int *pgs, int *po, int *pst, int
         break;
       }
       case STRSXP: {
-        SEXP *px = STRING_PTR(x), *pres = STRING_PTR(res);
+        SEXP *px = SEXPPTR(x), *pres = SEXPPTR(res);
         #pragma omp parallel for num_threads(nthreads)
         for(int gr = 0; gr < ng; ++gr)
           pres[gr] = pgs[gr] == 0 ? NA_STRING : w_mode_string(px + pst[gr]-1, pw + pst[gr]-1, po, pgs[gr], 1, narm, ret);
@@ -843,7 +843,7 @@ SEXP w_mode_g_impl(SEXP x, double *pw, int ng, int *pgs, int *po, int *pst, int
         break;
       }
       case STRSXP: {
-        SEXP *px = STRING_PTR(x), *pres = STRING_PTR(res);
+        SEXP *px = SEXPPTR(x), *pres = SEXPPTR(res);
         #pragma omp parallel for num_threads(nthreads)
         for(int gr = 0; gr < ng; ++gr)
           pres[gr] = pgs[gr] == 0 ? NA_STRING : w_mode_string(px, pw, po + pst[gr]-1, pgs[gr], 0, narm, ret);
@@ -1039,7 +1039,7 @@ SEXP fmodemC(SEXP x, SEXP g, SEXP w, SEXP Rnarm, SEXP Rdrop, SEXP Rret, SEXP Rnt
         break;
       }
       case STRSXP: {
-        SEXP *px = STRING_PTR(x), *restrict pres = STRING_PTR(res);
+        SEXP *px = SEXPPTR(x), *restrict pres = SEXPPTR(res);
         if(nullw) {
           #pragma omp parallel for num_threads(nthreads)
           for(int j = 0; j < col; ++j) pres[j] = mode_string(px + j*l, &l, l, 1, narm, ret);
@@ -1136,7 +1136,7 @@ SEXP fmodemC(SEXP x, SEXP g, SEXP w, SEXP Rnarm, SEXP Rdrop, SEXP Rret, SEXP Rnt
         break;
       }
       case STRSXP: {
-        SEXP *px = STRING_PTR(x), *restrict pres = STRING_PTR(res);
+        SEXP *px = SEXPPTR(x), *restrict pres = SEXPPTR(res);
         if(nullw) {
           #pragma omp parallel for num_threads(nthreads)
           for(int j = 0; j < col; ++j) {
@@ -1212,7 +1212,7 @@ SEXP fmodemC(SEXP x, SEXP g, SEXP w, SEXP Rnarm, SEXP Rdrop, SEXP Rret, SEXP Rnt
         break;
       }
       case STRSXP: {
-        SEXP *px = STRING_PTR(x), *restrict pres = STRING_PTR(res);
+        SEXP *px = SEXPPTR(x), *restrict pres = SEXPPTR(res);
         if(nullw) {
           #pragma omp parallel for num_threads(nthreads)
           for(int j = 0; j < col; ++j) {
diff --git a/src/fndistinct.c b/src/fndistinct.c
index 1d48b388..4298638b 100644
--- a/src/fndistinct.c
+++ b/src/fndistinct.c
@@ -242,7 +242,7 @@ int ndistinct_impl_int(SEXP x, int narm) {
       return isFactor(x) ? ndistinct_fct(INTEGER(x), &l, l, nlevels(x), 1, narm) :
                            ndistinct_int(INTEGER(x), &l, l, 1, narm);
     case LGLSXP: return ndistinct_logi(LOGICAL(x), &l, l, 1, narm);
-    case STRSXP: return ndistinct_string(STRING_PTR(x), &l, l, 1, narm);
+    case STRSXP: return ndistinct_string(SEXPPTR(x), &l, l, 1, narm);
     default: error("Not Supported SEXP Type: '%s'", type2char(TYPEOF(x)));
   }
 }
@@ -291,7 +291,7 @@ SEXP ndistinct_g_impl(SEXP x, const int ng, const int *restrict pgs, const int *
         break;
       }
       case STRSXP: {
-        const SEXP *px = STRING_PTR(x);
+        const SEXP *px = SEXPPTR(x);
         #pragma omp parallel for num_threads(nthreads)
         for(int gr = 0; gr < ng; ++gr)
           pres[gr] = pgs[gr] == 0 ? 0 : ndistinct_string(px + pst[gr]-1, po, pgs[gr], 1, narm);
@@ -330,7 +330,7 @@ SEXP ndistinct_g_impl(SEXP x, const int ng, const int *restrict pgs, const int *
         break;
       }
       case STRSXP: {
-        const SEXP *px = STRING_PTR(x);
+        const SEXP *px = SEXPPTR(x);
         #pragma omp parallel for num_threads(nthreads)
         for(int gr = 0; gr < ng; ++gr)
           pres[gr] = pgs[gr] == 0 ? 0 : ndistinct_string(px, po + pst[gr]-1, pgs[gr], 0, narm);
@@ -484,7 +484,7 @@ SEXP fndistinctmC(SEXP x, SEXP g, SEXP Rnarm, SEXP Rdrop, SEXP Rnthreads) {
         break;
       }
       case STRSXP: {
-        SEXP *px = STRING_PTR(x);
+        SEXP *px = SEXPPTR(x);
         #pragma omp parallel for num_threads(nthreads)
         for(int j = 0; j < col; ++j)
           pres[j] = ndistinct_string(px + j*l, &l, l, 1, narm);
@@ -555,7 +555,7 @@ SEXP fndistinctmC(SEXP x, SEXP g, SEXP Rnarm, SEXP Rdrop, SEXP Rnthreads) {
           break;
         }
         case STRSXP: {
-          SEXP *px = STRING_PTR(x);
+          SEXP *px = SEXPPTR(x);
           #pragma omp parallel for num_threads(nthreads)
           for(int j = 0; j < col; ++j) {
             int jng = j * ng;
@@ -602,7 +602,7 @@ SEXP fndistinctmC(SEXP x, SEXP g, SEXP Rnarm, SEXP Rdrop, SEXP Rnthreads) {
           break;
         }
         case STRSXP: {
-          SEXP *px = STRING_PTR(x);
+          SEXP *px = SEXPPTR(x);
           #pragma omp parallel for num_threads(nthreads)
           for(int j = 0; j < col; ++j) {
             int jng = j * ng;
diff --git a/src/fnobs.c b/src/fnobs.c
index e1500bcc..75c3427e 100644
--- a/src/fnobs.c
+++ b/src/fnobs.c
@@ -18,7 +18,7 @@ SEXP fnobsC(SEXP x, SEXP Rng, SEXP g) {
         break;
       }
       case STRSXP: {
-        SEXP *px = STRING_PTR(x);
+        SEXP *px = SEXPPTR(x);
         for(int i = 0; i != l; ++i) if(px[i] != NA_STRING) ++n;
         break;
       }
@@ -48,7 +48,7 @@ SEXP fnobsC(SEXP x, SEXP Rng, SEXP g) {
         break;
       }
       case STRSXP: {
-        SEXP *px = STRING_PTR(x);
+        SEXP *px = SEXPPTR(x);
         for(int i = 0; i != l; ++i) if(px[i] != NA_STRING) ++pn[pg[i]];
         break;
       }
@@ -102,7 +102,7 @@ SEXP fnobsmC(SEXP x, SEXP Rng, SEXP g, SEXP Rdrop) {
         break;
       }
       case STRSXP: {
-        SEXP *px = STRING_PTR(x);
+        SEXP *px = SEXPPTR(x);
         for(int j = 0; j != col; ++j) {
           int nj = 0, end = l * j + l;
           for(int i = l * j; i != end; ++i) if(px[i] != NA_STRING) ++nj;
@@ -145,7 +145,7 @@ SEXP fnobsmC(SEXP x, SEXP Rng, SEXP g, SEXP Rdrop) {
         break;
       }
       case STRSXP: {
-        SEXP *px = STRING_PTR(x)-l;
+        SEXP *px = SEXPPTR(x)-l;
         for(int j = 0; j != col; ++j) {
           pn += ng; px += l;
           for(int i = 0; i != l; ++i) if(px[i] != NA_STRING) ++pn[pg[i]];
diff --git a/src/gsplit.c b/src/gsplit.c
index 860e925e..41041e5f 100644
--- a/src/gsplit.c
+++ b/src/gsplit.c
@@ -90,9 +90,9 @@ SEXP gsplit(SEXP x, SEXP gobj, SEXP toint) {
         break;
       }
       case STRSXP: {
-        const SEXP *px = STRING_PTR(x);
+        const SEXP *px = SEXPPTR(x);
         for(int j = 0, gsj; j != ng; ++j) {
-          SEXP *pgj = STRING_PTR(pres[j]);
+          SEXP *pgj = SEXPPTR(pres[j]);
           gsj = pgs[j];
           for(int i = 0; i != gsj; ++i) pgj[i] = px[count++];
         }
@@ -160,9 +160,9 @@ SEXP gsplit(SEXP x, SEXP gobj, SEXP toint) {
         break;
       }
       case STRSXP: {
-        SEXP *px = STRING_PTR(x);
+        SEXP *px = SEXPPTR(x);
         for(int i = 0; i != ng; ++i) {
-          SEXP *pri = STRING_PTR(pres[i]);
+          SEXP *pri = SEXPPTR(pres[i]);
           for(int j = ps[i]-1, end = ps[i]+pgs[i]-1, k = 0; j < end; ++j) pri[k++] = px[po[j]-1];
         }
         break;
@@ -229,10 +229,10 @@ SEXP gsplit(SEXP x, SEXP gobj, SEXP toint) {
         break;
       }
       case STRSXP: {
-        const SEXP *px = STRING_PTR(x);
+        const SEXP *px = SEXPPTR(x);
         for(int i = 0, gi; i != l; ++i) {
           gi = pg[i]-1;
-          STRING_PTR(pres[gi])[count[gi]++] = px[i];
+          SEXPPTR(pres[gi])[count[gi]++] = px[i];
         }
         break;
       }
@@ -305,7 +305,7 @@ SEXP greorder(SEXP x, SEXP gobj) {
       break;
     }
     case STRSXP: {
-      SEXP *px = STRING_PTR(x), *pr = STRING_PTR(res);
+      SEXP *px = SEXPPTR(x), *pr = SEXPPTR(res);
       for(int i = 0, k = 0; i != ng; ++i) {
         for(int j = ps[i]-1, end = ps[i]+pgs[i]-1; j < end; ++j) pr[po[j]-1] = px[k++];
       }
@@ -355,7 +355,7 @@ SEXP greorder(SEXP x, SEXP gobj) {
       break;
     }
     case STRSXP: {
-      SEXP *px = STRING_PTR(x), *pr = STRING_PTR(res);
+      SEXP *px = SEXPPTR(x), *pr = SEXPPTR(res);
       for(int i = 0; i != l; ++i) pr[i] = px[cgs[pg[i]]+count[pg[i]]++];
       break;
     }
diff --git a/src/join.c b/src/join.c
index aa4894c2..ff087751 100644
--- a/src/join.c
+++ b/src/join.c
@@ -302,8 +302,8 @@ SEXP sort_merge_join(SEXP x, SEXP table, SEXP ot, SEXP count) {
         break;
       case STRSXP:
         checkEncodings(pci[0]); checkEncodings(pci[1]);
-        if(i == 0) sort_merge_join_string(STRING_PTR_RO(pci[0]), STRING_PTR_RO(pci[1])-1, pg, ptab, pot, nx, nt, pres);
-        else sort_merge_join_string_second(STRING_PTR_RO(pci[0]), STRING_PTR_RO(pci[1])-1, pg, ptab, pot, nx, nt, pres);
+        if(i == 0) sort_merge_join_string(SEXPPTR_RO(pci[0]), SEXPPTR_RO(pci[1])-1, pg, ptab, pot, nx, nt, pres);
+        else sort_merge_join_string_second(SEXPPTR_RO(pci[0]), SEXPPTR_RO(pci[1])-1, pg, ptab, pot, nx, nt, pres);
         break;
       case CPLXSXP:
         if(i == 0) sort_merge_join_complex(COMPLEX_RO(pci[0]), COMPLEX_RO(pci[1])-1, pg, ptab, pot, nx, nt, pres);
diff --git a/src/kit.h b/src/kit.h
index 97d51902..cf4e4103 100644
--- a/src/kit.h
+++ b/src/kit.h
@@ -12,7 +12,7 @@
 // #endif
 #include <Rinternals.h>
 // #include <stdlib.h>
-#include <stdint.h> // needed for intptr_t on linux
+#include <stdint.h> // needed for uintptr_t on linux
 // #include <stdbool.h>
 // #ifdef _OPENMP
 // #include<omp.h>
@@ -45,6 +45,7 @@
 // #define SEXP_T ScalarLogical(TRUE)
 #define NOGE(x, l) ((x < 0) || (x >= l))
 #define HASH(key, K)  (3141592653U * (unsigned int)(key) >> (32 - (K)))
+#define HASHK(key, K)  (3141592653U * (unsigned int)(key) >> (K))
 #define N_ISNAN(x, y) (!ISNAN(x) && !ISNAN(y))
 #define B_IsNA(x, y)  (R_IsNA(x) && R_IsNA(y))
 #define B_IsNaN(x, y) (R_IsNaN(x) && R_IsNaN(y))
@@ -54,6 +55,7 @@
 #define C_ISNAN(x, y) (B_ISNAN(x, y) || (N_ISNAN(x, y) && x == y))
 #define REQUAL(x, y)  (N_ISNAN(x, y) ? (x == y) : (B_IsNA(x, y) || B_IsNaN(x, y)))
 #define CEQUAL(x, y) ((N_ISNAN(x.r, x.i) && N_ISNAN(y.r, y.i)) ? (x.r == y.r && x.i == y.i) : (C_IsNA(x) ? C_IsNA(y) : (C_IsNA(y) ? 0 : (C_ISNAN(x.r, y.r) && C_ISNAN(x.i, y.i)))))
+#define SEXPPTR(x) ((SEXP *)DATAPTR(x))  // To replace STRING_PTR
 // #define STR_DF mkString("data.frame")
 // #define MAX(a,b) (((a)>(b))?(a):(b))
 // #define IS_LOGICAL(x) (isLogical(x) && LENGTH(x)==1)
diff --git a/src/kit_dup.c b/src/kit_dup.c
index 07a99953..22f6452a 100644
--- a/src/kit_dup.c
+++ b/src/kit_dup.c
@@ -175,7 +175,7 @@ SEXP dupVecIndex(SEXP x) {
     }
   } break;
   case STRSXP: {
-    const SEXP *restrict px = STRING_PTR(x);
+    const SEXP *restrict px = SEXPPTR(x);
     for (int i = 0; i != n; ++i) {
       id = HASH(((uintptr_t) px[i] & 0xffffffff), K);
       while(h[id]) {
@@ -338,7 +338,7 @@ SEXP dupVecIndexKeepNA(SEXP x) {
     }
   } break;
   case STRSXP: {
-    const SEXP *restrict px = STRING_PTR(x);
+    const SEXP *restrict px = SEXPPTR(x);
     for (int i = 0; i != n; ++i) {
       if(px[i] == NA_STRING) {
         pans_i[i] = NA_INTEGER;
@@ -452,7 +452,7 @@ SEXP dupVecIndexTwoVectors(SEXP x, SEXP y) {
           }
         } break;
         case STRSXP: {
-          const SEXP *restrict px = STRING_PTR_RO(x), *restrict py = STRING_PTR_RO(y);
+          const SEXP *restrict px = SEXPPTR_RO(x), *restrict py = SEXPPTR_RO(y);
           for (int i = 0; i != n; ++i) {
             id = HASH(64988430769U * ((uintptr_t)px[i] & 0xffffffff) + ((uintptr_t)py[i] & 0xffffffff), K); // Best combination it seems
             while(h[id]) {
@@ -513,7 +513,7 @@ SEXP dupVecIndexTwoVectors(SEXP x, SEXP y) {
         }
         // Second case: real and string
       } else if ((tx == REALSXP && ty == STRSXP) || (tx == STRSXP && ty == REALSXP)) {
-        const SEXP *restrict ps = STRING_PTR_RO(tx == STRSXP ? x : y);
+        const SEXP *restrict ps = SEXPPTR_RO(tx == STRSXP ? x : y);
         const double *restrict pr = REAL_RO(tx == REALSXP ? x : y);
         union uno tpv;
         for (int i = 0; i != n; ++i) {
@@ -534,7 +534,7 @@ SEXP dupVecIndexTwoVectors(SEXP x, SEXP y) {
         // Third case: integer and string
       } else if((tx == INTSXP && ty == STRSXP) || (tx == STRSXP && ty == INTSXP)) {
         const int *restrict pi = INTEGER_RO(tx == INTSXP ? x : y);
-        const SEXP *restrict ps = STRING_PTR_RO(tx == STRSXP ? x : y);
+        const SEXP *restrict ps = SEXPPTR_RO(tx == STRSXP ? x : y);
         for (int i = 0; i != n; ++i) {
           id = HASH(pi[i] * ((uintptr_t)ps[i] & 0xffffffff), K);
           while(h[id]) {
@@ -702,7 +702,7 @@ int dupVecSecond(int *restrict pidx, int *restrict pans_i, SEXP x, const int n,
     }
   } break;
   case STRSXP: {
-    const SEXP *restrict px = STRING_PTR(x);
+    const SEXP *restrict px = SEXPPTR(x);
     const unsigned int mult = (M-1) / ng; // -1 because C is zero indexed
     for (int i = 0; i != n; ++i) {
       id = (pidx[i]*mult) ^ HASH(((uintptr_t) px[i] & 0xffffffff), K); // HASH(((uintptr_t) px[i] & 0xffffffff) ^ pidx[i], K) + pidx[i];
@@ -984,7 +984,7 @@ SEXP funiqueC(SEXP x) {
     for(int i = 0; i != g; ++i) pres[i] = px[st[i]];
   } break;
   case STRSXP: {
-    const SEXP *restrict px = STRING_PTR(x);
+    const SEXP *restrict px = SEXPPTR(x);
     for (int i = 0; i != n; ++i) {
       id = HASH(((uintptr_t) px[i] & 0xffffffff), K);
       while(h[id]) {
@@ -1001,7 +1001,7 @@ SEXP funiqueC(SEXP x) {
       return x;
     }
     PROTECT(res = allocVector(STRSXP, g)); ++nprotect;
-    SEXP *restrict pres = STRING_PTR(res);
+    SEXP *restrict pres = SEXPPTR(res);
     for(int i = 0; i != g; ++i) pres[i] = px[st[i]];
   } break;
   }
diff --git a/src/match.c b/src/match.c
index 2fc97489..c0d8f5e8 100644
--- a/src/match.c
+++ b/src/match.c
@@ -250,7 +250,7 @@ SEXP match_single(SEXP x, SEXP table, SEXP nomatch) {
     }
   } break;
   case STRSXP: {
-    const SEXP *restrict px = STRING_PTR(x), *restrict pt = STRING_PTR(table);
+    const SEXP *restrict px = SEXPPTR(x), *restrict pt = SEXPPTR(table);
     // fill hash table with indices of 'table'
     for (int i = 0; i != nt; ++i) {
       id = HASH(((uintptr_t) pt[i] & 0xffffffff), K);
@@ -428,8 +428,8 @@ SEXP match_two_vectors(SEXP x, SEXP table, SEXP nomatch) {
         }
       } break;
       case STRSXP: {
-        const SEXP *restrict px1 = STRING_PTR(pc1[0]), *restrict px2 = STRING_PTR(pc2[0]),
-                   *restrict pt1 = STRING_PTR(pc1[1]), *restrict pt2 = STRING_PTR(pc2[1]);
+        const SEXP *restrict px1 = SEXPPTR(pc1[0]), *restrict px2 = SEXPPTR(pc2[0]),
+                   *restrict pt1 = SEXPPTR(pc1[1]), *restrict pt2 = SEXPPTR(pc2[1]);
         // fill hash table with indices of 'table'
         for (int i = 0; i != nt; ++i) {
           id = HASH(64988430769U * ((uintptr_t)pt1[i] & 0xffffffff) + ((uintptr_t)pt2[i] & 0xffffffff), K);
@@ -523,7 +523,7 @@ SEXP match_two_vectors(SEXP x, SEXP table, SEXP nomatch) {
     } else if ((t1 == REALSXP && t2 == STRSXP) || (t1 == STRSXP && t2 == REALSXP)) {
       const int rev = t1 == STRSXP;
       const double *restrict pxr = REAL(VECTOR_ELT(pc[rev], 0)), *restrict ptr = REAL(VECTOR_ELT(pc[rev], 1));
-      const SEXP *restrict pxs = STRING_PTR(VECTOR_ELT(pc[1-rev], 0)), *restrict pts = STRING_PTR(VECTOR_ELT(pc[1-rev], 1));
+      const SEXP *restrict pxs = SEXPPTR(VECTOR_ELT(pc[1-rev], 0)), *restrict pts = SEXPPTR(VECTOR_ELT(pc[1-rev], 1));
       union uno tpv;
       // fill hash table with indices of 'table'
       for (int i = 0; i != nt; ++i) {
@@ -554,7 +554,7 @@ SEXP match_two_vectors(SEXP x, SEXP table, SEXP nomatch) {
     } else if((t1 == INTSXP && t2 == STRSXP) || (t1 == STRSXP && t2 == INTSXP)) {
       const int rev = t1 == STRSXP;
       const int *restrict pxi = INTEGER(VECTOR_ELT(pc[rev], 0)), *restrict pti = INTEGER(VECTOR_ELT(pc[rev], 1));
-      const SEXP *restrict pxs = STRING_PTR(VECTOR_ELT(pc[1-rev], 0)), *restrict pts = STRING_PTR(VECTOR_ELT(pc[1-rev], 1));
+      const SEXP *restrict pxs = SEXPPTR(VECTOR_ELT(pc[1-rev], 0)), *restrict pts = SEXPPTR(VECTOR_ELT(pc[1-rev], 1));
 
       // fill hash table with indices of 'table'
       for (int i = 0; i != nt; ++i) {
@@ -643,8 +643,8 @@ void match_two_vectors_extend(const SEXP *pc, const int nmv, const int n, const
       }
     } break;
     case STRSXP: {
-      const SEXP *restrict px1 = STRING_PTR(pc1[0]), *restrict px2 = STRING_PTR(pc2[0]),
-                 *restrict pt1 = STRING_PTR(pc1[1]), *restrict pt2 = STRING_PTR(pc2[1]);
+      const SEXP *restrict px1 = SEXPPTR(pc1[0]), *restrict px2 = SEXPPTR(pc2[0]),
+                 *restrict pt1 = SEXPPTR(pc1[1]), *restrict pt2 = SEXPPTR(pc2[1]);
       // fill hash table with indices of 'table'
       for (int i = 0; i != nt; ++i) {
         id = HASH(64988430769U * ((uintptr_t)pt1[i] & 0xffffffff) + ((uintptr_t)pt2[i] & 0xffffffff), K);
@@ -747,7 +747,7 @@ void match_two_vectors_extend(const SEXP *pc, const int nmv, const int n, const
     } else if ((t1 == REALSXP && t2 == STRSXP) || (t1 == STRSXP && t2 == REALSXP)) {
       const int rev = t1 == STRSXP;
       const double *restrict pxr = REAL(VECTOR_ELT(pc[rev], 0)), *restrict ptr = REAL(VECTOR_ELT(pc[rev], 1));
-      const SEXP *restrict pxs = STRING_PTR(VECTOR_ELT(pc[1-rev], 0)), *restrict pts = STRING_PTR(VECTOR_ELT(pc[1-rev], 1));
+      const SEXP *restrict pxs = SEXPPTR(VECTOR_ELT(pc[1-rev], 0)), *restrict pts = SEXPPTR(VECTOR_ELT(pc[1-rev], 1));
       union uno tpv;
       // fill hash table with indices of 'table'
       for (int i = 0; i != nt; ++i) {
@@ -781,7 +781,7 @@ void match_two_vectors_extend(const SEXP *pc, const int nmv, const int n, const
     } else if((t1 == INTSXP && t2 == STRSXP) || (t1 == STRSXP && t2 == INTSXP)) {
       const int rev = t1 == STRSXP;
       const int *restrict pxi = INTEGER(VECTOR_ELT(pc[rev], 0)), *restrict pti = INTEGER(VECTOR_ELT(pc[rev], 1));
-      const SEXP *restrict pxs = STRING_PTR(VECTOR_ELT(pc[1-rev], 0)), *restrict pts = STRING_PTR(VECTOR_ELT(pc[1-rev], 1));
+      const SEXP *restrict pxs = SEXPPTR(VECTOR_ELT(pc[1-rev], 0)), *restrict pts = SEXPPTR(VECTOR_ELT(pc[1-rev], 1));
 
       // fill hash table with indices of 'table'
       for (int i = 0; i != nt; ++i) {
@@ -871,7 +871,7 @@ void match_additional(const SEXP *pcj, const int nmv, const int n, const int nt,
       }
     } break;
     case STRSXP: {
-      const SEXP *restrict px = STRING_PTR(pcj[0]), *restrict pt = STRING_PTR(pcj[1]);
+      const SEXP *restrict px = SEXPPTR(pcj[0]), *restrict pt = SEXPPTR(pcj[1]);
       // fill hash table with indices of 'table'
       for (int i = 0; i != nt; ++i) {
         if(ptab_copy[i] == nmv) {
@@ -964,7 +964,7 @@ void match_rest(const SEXP *pcj, const int nmv, const int n, const int nt, int *
       }
     } break;
     case STRSXP: {
-      const SEXP *restrict px = STRING_PTR(pcj[0]), *restrict pt = STRING_PTR(pcj[1])-1;
+      const SEXP *restrict px = SEXPPTR(pcj[0]), *restrict pt = SEXPPTR(pcj[1])-1;
       for (int i = 0; i != n; ++i) {
         if(pans[i] == nmv) continue;
         if(px[i] != pt[pans[i]]) pans[i] = nmv;
diff --git a/src/pivot.c b/src/pivot.c
index 4934000f..a3a61c8a 100644
--- a/src/pivot.c
+++ b/src/pivot.c
@@ -119,7 +119,7 @@ SEXP pivot_long(SEXP data, SEXP ind, SEXP idcol) {
         pid += end; ++v;
       }
     } else {
-      SEXP *restrict pid = STRING_PTR(id_column), *pnam = STRING_PTR(names);
+      SEXP *restrict pid = SEXPPTR(id_column), *pnam = SEXPPTR(names);
       for (int j = 0, end = 0; j != l; ++j) {
         SEXP namj = pnam[j];
         end = length(pind[j]); // SIMD??
@@ -314,9 +314,9 @@ SEXP pivot_wide(SEXP index, SEXP id, SEXP column, SEXP fill, SEXP Rnthreads, SEX
       break;
     }
     case STRSXP: {
-      const SEXP *restrict pc = STRING_PTR_RO(column);
+      const SEXP *restrict pc = SEXPPTR_RO(column);
       if(aggfun > 3) error("Cannot aggregate character column with sum, mean, min, or max.");
-      AGGFUN_SWITCH_CAT(STRING_PTR, pc[i] != NA_STRING);
+      AGGFUN_SWITCH_CAT(SEXPPTR, pc[i] != NA_STRING);
       break;
     }
     case VECSXP:
diff --git a/src/programming.c b/src/programming.c
index aca30e4f..8b88229f 100644
--- a/src/programming.c
+++ b/src/programming.c
@@ -31,11 +31,11 @@ SEXP Cna_rm(SEXP x) {
     return out;
   }
   case STRSXP: {
-    const SEXP *xd = STRING_PTR(x);
+    const SEXP *xd = SEXPPTR(x);
     for (int i = 0; i != n; ++i) if(xd[i] == NA_STRING) ++k;
     if(k == 0) return x;
     SEXP out = PROTECT(allocVector(STRSXP, n - k));
-    SEXP *pout = STRING_PTR(out);
+    SEXP *pout = SEXPPTR(out);
     k = 0;
     for (int i = 0; i != n; ++i) if(xd[i] != NA_STRING) pout[k++] = xd[i];
     copyMostAttrib(x, out);
@@ -60,7 +60,7 @@ SEXP Cna_rm(SEXP x) {
 
 // Helper function to find a single string in factor levels
 int fchmatch(SEXP x, SEXP val, int nomatch) {
-  const SEXP *px = STRING_PTR(x), v = asChar(val);
+  const SEXP *px = SEXPPTR(x), v = asChar(val);
   for(int i = 0, l = length(x); i != l; ++i) if(px[i] == v) return i + 1;
   return nomatch;
 }
@@ -109,8 +109,8 @@ if(length(val) == n && n > 1) {
   }
   case STRSXP:
   {
-    const SEXP *px = STRING_PTR(x);
-    const SEXP *pv = STRING_PTR(val);
+    const SEXP *px = SEXPPTR(x);
+    const SEXP *pv = SEXPPTR(val);
     WHICHVLOOPLX
     break;
   }
@@ -155,7 +155,7 @@ if(length(val) == n && n > 1) {
   }
   case STRSXP:
   {
-    const SEXP *px = STRING_PTR(x);
+    const SEXP *px = SEXPPTR(x);
     const SEXP v = PROTECT(asChar(val));
     WHICHVLOOP
     UNPROTECT(1);
@@ -216,7 +216,7 @@ case REALSXP:
 }
 case STRSXP:
 {
-  const SEXP *px = STRING_PTR(x);
+  const SEXP *px = SEXPPTR(x);
   const SEXP v = asChar(val);
   ALLANYVLOOP
   break;
@@ -387,7 +387,7 @@ SEXP setcopyv(SEXP x, SEXP val, SEXP rep, SEXP Rinvert, SEXP Rset, SEXP Rind1) {
   }
   case STRSXP:
   {
-    SEXP *restrict px = set ? STRING_PTR(x) : STRING_PTR(ans);
+    SEXP *restrict px = set ? SEXPPTR(x) : SEXPPTR(ans);
     if(lv == 1 && ind1 == 0) {
       const SEXP v = PROTECT(asChar(val));
       if(lr == 1) {
@@ -395,7 +395,7 @@ SEXP setcopyv(SEXP x, SEXP val, SEXP rep, SEXP Rinvert, SEXP Rset, SEXP Rind1) {
         setcopyvLOOP(r)
         UNPROTECT(1);
       } else {
-        const SEXP *restrict pr = STRING_PTR(rep);
+        const SEXP *restrict pr = SEXPPTR(rep);
         setcopyvLOOP(pr[i])
       }
       UNPROTECT(1);
@@ -406,7 +406,7 @@ SEXP setcopyv(SEXP x, SEXP val, SEXP rep, SEXP Rinvert, SEXP Rset, SEXP Rind1) {
         setcopyvLOOPLVEC1
         UNPROTECT(1);
       } else {
-        const SEXP *restrict pr = STRING_PTR(rep);
+        const SEXP *restrict pr = SEXPPTR(rep);
         setcopyvLOOPLVEC
       }
     }
@@ -760,7 +760,7 @@ SEXP na_locf(SEXP x, SEXP Rset) {
   }
   case STRSXP:
   {
-    SEXP *data = STRING_PTR(x);
+    SEXP *data = SEXPPTR(x);
     SEXP last = data[0];
     for (int i = 0; i < n; i++) {
       if (data[i] == NA_STRING) {
@@ -826,7 +826,7 @@ SEXP na_focb(SEXP x, SEXP Rset) {
   }
   case STRSXP:
   {
-    SEXP *data = STRING_PTR(x);
+    SEXP *data = SEXPPTR(x);
     SEXP last = data[0];
     for (int i = n; i--; ) {
       if (data[i] == NA_STRING) {
diff --git a/src/small_helper.c b/src/small_helper.c
index 8c4e8731..8b8d1d22 100644
--- a/src/small_helper.c
+++ b/src/small_helper.c
@@ -66,7 +66,7 @@ SEXP falloc(SEXP value, SEXP n, SEXP simplify)  {
       break;
     }
     case STRSXP: {
-      SEXP val = asChar(value), *pout = STRING_PTR(out);
+      SEXP val = asChar(value), *pout = SEXPPTR(out);
       for(int i = 0; i != l; ++i) pout[i] = val;
       break;
     }
@@ -169,7 +169,7 @@ SEXP multiassign(SEXP lhs, SEXP rhs, SEXP envir) {
     return R_NilValue;
   }
   if(length(rhs) != n) error("length(lhs) must be equal to length(rhs)");
-  SEXP *plhs = STRING_PTR(lhs);
+  SEXP *plhs = SEXPPTR(lhs);
   switch(TYPEOF(rhs)) { // installTrChar translates to native encoding, installChar does the same now, but also is available on older systems.
     case REALSXP: {
       double *prhs = REAL(rhs);
@@ -188,7 +188,7 @@ SEXP multiassign(SEXP lhs, SEXP rhs, SEXP envir) {
       break;
     }
     case STRSXP: {
-      SEXP *prhs = STRING_PTR(rhs);
+      SEXP *prhs = SEXPPTR(rhs);
       for(int i = 0; i < n; ++i) {
         SEXP nam = installChar(plhs[i]);
         defineVar(nam, ScalarString(prhs[i]), envir);
@@ -235,7 +235,7 @@ SEXP vlabels(SEXP x, SEXP attrn, SEXP usenam) {
     return labx;
   }
   SEXP res = PROTECT(allocVector(STRSXP, l));
-  SEXP *pres = STRING_PTR(res);
+  SEXP *pres = SEXPPTR(res);
   const SEXP *px = SEXPPTR_RO(x);
   for(int i = 0; i < l; ++i) {
     SEXP labxi = getAttrib(px[i], sym_attrn);