fread nul in the middle of a field too

inst/tests/tests.Rraw

@@ -14033,8 +14033,11 @@ for (nNUL in 0:3) {
   test(2025.08, fread(f, skip=1, header=TRUE), ans)
   test(2025.09, fread(f), ans)
 }
-writeBin(c(charToRaw("A,B,C\n1,foo,5\n2,"), as.raw(0),charToRaw("bar"), as.raw(0),as.raw(0), charToRaw(",6\n")), con=f)
+makeNul = function(str){ tt=charToRaw(str); tt[tt==42L]=as.raw(0); writeBin(tt, con=f)}  # "*" (42) represents NUL
+makeNul("A,B,C\n1,foo,5\n2,*bar**,6\n")
 test(2025.10, fread(f), data.table(A=1:2, B=c("foo","bar"), C=5:6))
+makeNul('A,B,C\n1,foo*bar,3\n2,**"**b*az*",4\n')
+test(2025.11, fread(f), data.table(A=1:2, B=c("foobar","baz"), C=3:4))
 
 # printing timezone, #2842
 DT = data.table(t1 = as.POSIXct("1982-04-26 13:34:56", tz = "Europe/Madrid"),t2 = as.POSIXct("2019-01-01 19:00:01",tz = "UTC"))

src/fread.c

@@ -497,7 +497,7 @@ static void Field(FieldParseContext *ctx)
   const char *fieldStart=ch;
   if (*ch!=quote || quoteRule==3) {
     // Most common case. Unambiguously not quoted. Simply search for sep|eol. If field contains sep|eol then it should have been quoted and we do not try to heal that.
-    while(!end_of_field(ch)) ch++;  // sep, \r, \n or \0 will end
+    while(!end_of_field(ch)) ch++;  // sep, \r, \n or eof will end
     *(ctx->ch) = ch;
     int fieldLen = (int)(ch-fieldStart);
     //if (stripWhite) {   // TODO:  do this if and the next one together once in bulk afterwards before push
@@ -516,15 +516,15 @@ static void Field(FieldParseContext *ctx)
   fieldStart++;  // step over opening quote
   switch(quoteRule) {
   case 0:  // quoted with embedded quotes doubled; the final unescaped " must be followed by sep|eol
-    while (*++ch) {
+    while (*++ch || ch<eof) {
       if (*ch==quote) {
         if (ch[1]==quote) { ch++; continue; }
         break;  // found undoubled closing quote
       }
     }
     break;
   case 1:  // quoted with embedded quotes escaped; the final unescaped " must be followed by sep|eol
-    while (*++ch) {
+    while (*++ch || ch<eof) {
       if (*ch=='\\' && (ch[1]==quote || ch[1]=='\\')) { ch++; continue; }
       if (*ch==quote) break;
     }
@@ -538,14 +538,14 @@ static void Field(FieldParseContext *ctx)
     // Under this rule, no eol may occur inside fields.
     {
       const char *ch2 = ch;
-      while (*++ch && *ch!='\n' && *ch!='\r') {
+      while ((*++ch || ch<eof) && *ch!='\n' && *ch!='\r') {
         if (*ch==quote && end_of_field(ch+1)) {ch2=ch; break;}  // (*1) regular ", ending; leave *ch on closing quote
         if (*ch==sep) {
           // first sep in this field
           // if there is a ", afterwards but before the next \n, use that; the field was quoted and it's still case (i) above.
           // Otherwise break here at this first sep as it's case (ii) above (the data contains a quote at the start and no sep)
           ch2 = ch;
-          while (*++ch2 && *ch2!='\n' && *ch2!='\r') {
+          while ((*++ch2 || ch2<eof) && *ch2!='\n' && *ch2!='\r') {
             if (*ch2==quote && end_of_field(ch2+1)) {
               ch = ch2;                                          // (*2) move on to that first ", -- that's this field's ending
               break;
@@ -568,7 +568,7 @@ static void Field(FieldParseContext *ctx)
     *(ctx->ch) = ch;
   } else {
     *(ctx->ch) = ch;
-    if (*ch=='\0' && quoteRule!=2) { target->off--; target->len++; }   // test 1324 where final field has open quote but not ending quote; include the open quote like quote rule 2
+    if (ch==eof && quoteRule!=2) { target->off--; target->len++; }   // test 1324 where final field has open quote but not ending quote; include the open quote like quote rule 2
     while(target->len>0 && ((ch[-1]==' ' && stripWhite) || ch[-1]=='\0')) { target->len--; ch--; }  // test 1551.6; trailing whitespace in field [67,V37] == "\"\"A\"\" ST       "
   }
 }

src/freadR.c

@@ -427,10 +427,24 @@ void pushBuffer(ThreadLocalFreadParsingContext *ctx)
           lenOff *source = buff8_lenoffs + off8;
           for (int i=0; i<nRows; i++) {
             int strLen = source->len;
-            if (strLen) {
+            if (strLen<=0) {
               // stringLen == INT_MIN => NA, otherwise not a NAstring was checked inside fread_mean
-              SET_STRING_ELT(dest, DTi+i, strLen<0 ? NA_STRING : mkCharLenCE(anchor + source->off, strLen, ienc));
-            } // else dest was already initialized with R_BlankString by allocVector()
+              if (strLen<0) SET_STRING_ELT(dest, DTi+i, NA_STRING); // else leave the "" in place that was initialized by allocVector()
+            } else {
+              const char *str = anchor + source->off;
+              int c=0;
+              while (c<strLen && str[c]) c++;
+              if (c<strLen) {
+                // embedded nul found; any at the beginning or the end of the field should have already been excluded but this will strip those too if present just in case
+                char *last = (char *)str+c;    // obtain write access to (const char *)anchor;
+                while (c<strLen) {
+                  if (str[c]) *last++=str[c];  // cow page write: saves allocation and management of a temp that would need to thread-safe in future.
+                  c++;                         //   This is only thread accessing this region. For non-mmap direct input nul are not possible (R would not have accepted nul earlier).
+                }
+                strLen = last-str;
+              }
+              SET_STRING_ELT(dest, DTi+i, mkCharLenCE(str, strLen, ienc));
+            }
             source += cnt8;
           }
           done++; // if just one string col near the start, don't loop over the other 10,000 cols. TODO? start on first too

src/nafill.c

@@ -41,7 +41,7 @@ SEXP colnamesInt(SEXP x, SEXP cols) {
 }
 
 void nafillDouble(double *x, uint_fast64_t nx, unsigned int type, double fill, ans_t *ans, bool verbose) {
-  double tic;
+  double tic=0.0;
   if (verbose) tic = omp_get_wtime();
   if (type==0) { // const
     for (uint_fast64_t i=0; i<nx; i++) {
@@ -62,7 +62,7 @@ void nafillDouble(double *x, uint_fast64_t nx, unsigned int type, double fill, a
 }
 
 void nafillInteger(int32_t *x, uint_fast64_t nx, unsigned int type, int32_t fill, ans_t *ans, bool verbose) {
-  double tic;
+  double tic=0.0;
   if (verbose) tic = omp_get_wtime();
   if (type==0) { // const
     for (uint_fast64_t i=0; i<nx; i++) {
@@ -161,7 +161,7 @@ SEXP nafillR(SEXP obj, SEXP type, SEXP fill, SEXP inplace, SEXP cols, SEXP verbo
     }
   }
 
-  double tic, toc;
+  double tic=0.0, toc=0.0;
   if (bverbose) tic = omp_get_wtime();
   #pragma omp parallel for if (nx>1) schedule(auto) num_threads(getDTthreads())
   for (R_len_t i=0; i<nx; i++) {