added nan parameter to fcoalesce

Author: Mukulyadav2004

R/wrappers.R

@@ -2,8 +2,8 @@
 # Very small (e.g. one line) R functions that just call C.
 # One file wrappers.R to avoid creating lots of small .R files.
 
-fcoalesce   = function(...) .Call(Ccoalesce, list(...), FALSE)
-setcoalesce = function(...) .Call(Ccoalesce, list(...), TRUE)
+fcoalesce   = function(..., nan=NA) .Call(Ccoalesce, list(...), FALSE, nan_is_na(nan))
+setcoalesce = function(..., nan=NA) .Call(Ccoalesce, list(...), TRUE, nan_is_na(nan))
 
 fifelse = function(test, yes, no, na=NA) .Call(CfifelseR, test, yes, no, na)
 fcase   = function(..., default=NA) {

man/coalesce.Rd

@@ -7,10 +7,11 @@ Fill in missing values in a vector by successively pulling from candidate vector
 Written in C, and multithreaded for numeric and factor types.
 }
 \usage{
-  fcoalesce(\dots)
+  fcoalesce(\dots, nan = NA)
 }
 \arguments{
   \item{\dots}{ A set of same-class vectors. These vectors can be supplied as separate arguments or as a single plain list, data.table or data.frame, see examples. }
+  \item{nan}{ (numeric vectors only) Either \code{NaN} or \code{NA}; if \code{NaN}, then \code{NaN} is treated as distinct from \code{NA}, otherwise they are treated the same during replacement. }
 }
 \details{
 Factor type is supported only when the factor levels of each item are equal.
@@ -31,6 +32,9 @@ z = c(11L, NA, 1L, 14L, NA, NA)
 fcoalesce(x, y, z)
 fcoalesce(list(x,y,z))   # same
 fcoalesce(x, list(y,z))  # same
+x_num = c(NaN, NA_real_, 3.0)
+fcoalesce(x_num, 1)           # default: NaN treated as missing -> c(1, 1, 3)
+fcoalesce(x_num, 1, nan=NaN)  # preserve NaN -> c(NaN, 1, 3)
 }
 \keyword{ data }
 

src/coalesce.c

@@ -6,10 +6,12 @@
     - The replacement of NAs with non-NA values from subsequent vectors
     - The conditional checks within parallelized loops
 */
-SEXP coalesce(SEXP x, SEXP inplaceArg) {
+SEXP coalesce(SEXP x, SEXP inplaceArg, SEXP nan_is_na_arg) {
   if (TYPEOF(x)!=VECSXP) internal_error(__func__, "input is list(...) at R level"); // # nocov
   if (!IS_TRUE_OR_FALSE(inplaceArg)) internal_error(__func__, "argument 'inplaceArg' must be TRUE or FALSE"); // # nocov
+  if (!IS_TRUE_OR_FALSE(nan_is_na_arg)) internal_error(__func__, "argument 'nan_is_na_arg' must be TRUE or FALSE"); // # nocov
   const bool inplace = LOGICAL(inplaceArg)[0];
+  const bool nan_is_na = LOGICAL(nan_is_na_arg)[0];
   const bool verbose = GetVerbose();
   int nprotect = 0;
   if (length(x)==0 || isNull(VECTOR_ELT(x,0))) return R_NilValue;  // coalesce(NULL, "foo") return NULL even though character type mismatches type NULL
@@ -106,7 +108,7 @@ SEXP coalesce(SEXP x, SEXP inplaceArg) {
         SEXP item = VECTOR_ELT(x, j+off);
         if (length(item)==1) {
           double tt = REAL(item)[0];
-          if (ISNAN(tt)) continue;
+          if (nan_is_na ? ISNAN(tt) : ISNA(tt)) continue;
           finalVal = tt;
           break;
         }
@@ -116,9 +118,9 @@ SEXP coalesce(SEXP x, SEXP inplaceArg) {
       #pragma omp parallel for num_threads(getDTthreads(nrow, true))
       for (int i=0; i<nrow; ++i) {
         double val=xP[i];
-        if (!ISNAN(val)) continue;
-        int j=0; while (ISNAN(val) && j<k) val=((double *)valP[j++])[i];
-        if (!ISNAN(val)) xP[i]=val; else if (final) xP[i]=finalVal;
+        if (!(nan_is_na ? ISNAN(val) : ISNA(val))) continue;
+        int j=0; while ((nan_is_na ? ISNAN(val) : ISNA(val)) && j<k) val=((double *)valP[j++])[i];
+        if (!(nan_is_na ? ISNAN(val) : ISNA(val))) xP[i]=val; else if (final) xP[i]=finalVal;
       }
     }
   } break;

src/data.table.h

@@ -251,7 +251,7 @@ SEXP nafillR(SEXP obj, SEXP type, SEXP fill, SEXP nan_is_na_arg, SEXP inplace, S
 SEXP between(SEXP x, SEXP lower, SEXP upper, SEXP incbounds, SEXP NAbounds, SEXP check);
 
 // coalesce.c
-SEXP coalesce(SEXP x, SEXP inplace);
+SEXP coalesce(SEXP x, SEXP inplace, SEXP nan_is_na_arg);
 
 // utils.c
 bool within_int32_repres(double x);