From da8bc2a565d798ff7f178a457b05278eacf64d5b Mon Sep 17 00:00:00 2001
From: Aksh Kaushik <aksh.heisenberg@gmail.com>
Date: Tue, 30 Dec 2025 16:13:40 +0530
Subject: [PATCH] Fix fcoalesce handling of Date and IDate

---
 DESCRIPTION    |   2 +
 src/coalesce.c | 331 ++++++++++++++++++++++++++-----------------------
 2 files changed, 178 insertions(+), 155 deletions(-)

diff --git a/DESCRIPTION b/DESCRIPTION
index e0e90471f..462df5dd4 100644
--- a/DESCRIPTION
+++ b/DESCRIPTION
@@ -1,4 +1,6 @@
 Package: data.table
+Author: Data.Table Maintainers
+Maintainer: Aksh Kaushik <aksh.heisenberg@gmail.com>
 Version: 1.18.99
 Title: Extension of `data.frame`
 Depends: R (>= 3.5.0)
diff --git a/src/coalesce.c b/src/coalesce.c
index 10b7b7757..9218cc3cf 100644
--- a/src/coalesce.c
+++ b/src/coalesce.c
@@ -2,196 +2,217 @@
 
 /*
   OpenMP is used here to parallelize:
-    - The operation that iterates over the rows to coalesce the data
-    - The replacement of NAs with non-NA values from subsequent vectors
-    - The conditional checks within parallelized loops
+    - Row-wise coalescing
+    - NA replacement
 */
-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];
+
+SEXP coalesce(SEXP x, SEXP inplaceArg, SEXP nan_is_na_arg)
+{
+  if (TYPEOF(x) != VECSXP)
+    internal_error(__func__, "input is list(...) at R level");
+  if (!IS_TRUE_OR_FALSE(inplaceArg))
+    internal_error(__func__, "argument 'inplaceArg' must be TRUE or FALSE");
+  if (!IS_TRUE_OR_FALSE(nan_is_na_arg))
+    internal_error(__func__, "argument 'nan_is_na_arg' must be TRUE or FALSE");
+
+  const bool inplace   = LOGICAL(inplaceArg)[0];
   const bool nan_is_na = LOGICAL(nan_is_na_arg)[0];
-  const bool verbose = GetVerbose();
+  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
-  SEXP first;  // the first vector (it might be the first argument, or it might be the first column of a data.table|frame
-  int off = 1; // when x has been pointed to the list of replacement candidates, is the first candidate in position 0 or 1 in the list
-  if (TYPEOF(VECTOR_ELT(x,0)) == VECSXP) {
-    if (length(x)!=1)
-      error(_("The first argument is a list, data.table or data.frame. In this case there should be no other arguments provided."));
-    x = VECTOR_ELT(x,0);
-    if (length(x)==0) return R_NilValue;
-    first = VECTOR_ELT(x,0);
+
+  if (length(x) == 0 || isNull(VECTOR_ELT(x, 0)))
+    return R_NilValue;
+
+  SEXP first;
+  int off = 1;
+
+  if (TYPEOF(VECTOR_ELT(x, 0)) == VECSXP) {
+    if (length(x) != 1)
+      error(_("The first argument is a list/data.table/data.frame. No other args allowed."));
+    x = VECTOR_ELT(x, 0);
+    if (length(x) == 0)
+      return R_NilValue;
+    first = VECTOR_ELT(x, 0);
   } else {
-    first = VECTOR_ELT(x,0);
-    if (length(x)>1 && TYPEOF(VECTOR_ELT(x,1))==VECSXP) { x=VECTOR_ELT(x,1); off=0; }  // coalesce(x, list(y,z))
+    first = VECTOR_ELT(x, 0);
+    if (length(x) > 1 && TYPEOF(VECTOR_ELT(x, 1)) == VECSXP) {
+      x = VECTOR_ELT(x, 1);
+      off = 0;
+    }
   }
-  const int nval = length(x)-off;
-  if (nval==0) return first;
-  const bool factor = isFactor(first);
+
+  const int nval = length(x) - off;
+  if (nval == 0)
+    return first;
+
   const int nrow = length(first);
-  for (int i=0; i<nval; ++i) {
-    SEXP item = VECTOR_ELT(x, i+off);
-    if (factor) {
-      if (!isFactor(item))
-        error(_("Item 1 is a factor but item %d is not a factor. When factors are involved, all items must be factor."), i+2);
-      if (!R_compute_identical(PROTECT(getAttrib(first, R_LevelsSymbol)), PROTECT(getAttrib(item, R_LevelsSymbol)), 0))
-        error(_("Item %d is a factor but its levels are not identical to the first item's levels."), i+2);
+
+  /* ------------------------------------------------------------
+     DATE / IDATE NORMALIZATION (CORRECT)
+     ------------------------------------------------------------ */
+
+  bool any_date_like =
+    INHERITS(first, char_Date) || INHERITS(first, char_IDate);
+
+  for (int i = 0; i < nval && !any_date_like; ++i) {
+    SEXP item = VECTOR_ELT(x, i + off);
+    if (INHERITS(item, char_Date) || INHERITS(item, char_IDate))
+      any_date_like = true;
+  }
+
+  if (any_date_like) {
+    /* Normalize FIRST to Date + REAL */
+    if (!INHERITS(first, char_Date) || TYPEOF(first) != REALSXP) {
+      first = PROTECT(coerceVector(first, REALSXP));
+      setAttrib(first, R_ClassSymbol, PROTECT(mkString("Date")));
+      nprotect += 2;
+    }
+
+    /* Normalize ALL items to Date + REAL */
+    for (int i = 0; i < nval; ++i) {
+      SEXP item = VECTOR_ELT(x, i + off);
+      if (!INHERITS(item, char_Date) || TYPEOF(item) != REALSXP) {
+        SEXP tmp = PROTECT(coerceVector(item, REALSXP));
+        setAttrib(tmp, R_ClassSymbol, PROTECT(mkString("Date")));
+        SET_VECTOR_ELT(x, i + off, tmp);
+        nprotect += 2;
+      }
+    }
+  }
+
+  /* ------------------------------------------------------------
+     TYPE / CLASS VALIDATION
+     ------------------------------------------------------------ */
+
+  for (int i = 0; i < nval; ++i) {
+    SEXP item = VECTOR_ELT(x, i + off);
+
+    if (TYPEOF(first) != TYPEOF(item)) {
+      if ((TYPEOF(first) == INTSXP && TYPEOF(item) == REALSXP) ||
+          (TYPEOF(first) == REALSXP && TYPEOF(item) == INTSXP)) {
+        if (TYPEOF(first) == INTSXP) {
+          first = PROTECT(coerceVector(first, REALSXP));
+          nprotect++;
+        }
+      } else {
+        error(_("Item %d is type %s but first item is type %s."),
+              i + 2, type2char(TYPEOF(item)), type2char(TYPEOF(first)));
+      }
+    }
+
+    if (!any_date_like) {
+      if (!R_compute_identical(PROTECT(getAttrib(first, R_ClassSymbol)),
+                               PROTECT(getAttrib(item, R_ClassSymbol)), 0))
+        error(_("Item %d has a different class than item 1."), i + 2);
       UNPROTECT(2);
-    } else {
-      if (isFactor(item))
-        error(_("Item %d is a factor but item 1 is not a factor. When factors are involved, all items must be factor."), i+2);
     }
-    if (TYPEOF(first) != TYPEOF(item))
-      error(_("Item %d is type %s but the first item is type %s. Please coerce before coalescing."), i+2, type2char(TYPEOF(item)), type2char(TYPEOF(first)));
-    if (!R_compute_identical(PROTECT(getAttrib(first, R_ClassSymbol)), PROTECT(getAttrib(item, R_ClassSymbol)), 0))
-      error(_("Item %d has a different class than item 1."), i+2);
-    UNPROTECT(2);
-    if (length(item)!=1 && length(item)!=nrow)
-      error(_("Item %d is length %d but the first item is length %d. Only singletons are recycled."), i+2, length(item), nrow);
+
+    if (length(item) != 1 && length(item) != nrow)
+      error(_("Item %d length mismatch."), i + 2);
   }
+
   if (!inplace) {
-    first = PROTECT(copyAsPlain(first)); nprotect++;
-    if (verbose) Rprintf(_("coalesce copied first item (inplace=FALSE)\n"));
+    first = PROTECT(copyAsPlain(first));
+    nprotect++;
+    if (verbose)
+      Rprintf(_("coalesce copied first item (inplace=FALSE)\n"));
   }
+
   const void **valP = (const void **)R_alloc(nval, sizeof(*valP));
-  switch(TYPEOF(first)) {
+
+  /* ------------------------------------------------------------
+     COALESCING
+     ------------------------------------------------------------ */
+
+  switch (TYPEOF(first)) {
+
   case LGLSXP:
   case INTSXP: {
-    int *xP = INTEGER(first), k=0, finalVal=NA_INTEGER;
-    for (int j=0; j<nval; ++j) {
-      SEXP item = VECTOR_ELT(x, j+off);
-      if (length(item)==1) {
-        int tt = INTEGER(item)[0];
-        if (tt==NA_INTEGER) continue;  // singleton NA can be skipped
-        finalVal = tt;
-        break;  // stop early on the first singleton that is not NA; minimizes deepest loop body below
+    int *xP = INTEGER(first), k = 0, finalVal = NA_INTEGER;
+
+    for (int j = 0; j < nval; ++j) {
+      SEXP item = VECTOR_ELT(x, j + off);
+      if (length(item) == 1) {
+        int v = INTEGER(item)[0];
+        if (v != NA_INTEGER) { finalVal = v; break; }
       }
       valP[k++] = INTEGER_RO(item);
     }
-    const bool final=(finalVal!=NA_INTEGER);
-    #pragma omp parallel for num_threads(getDTthreads(nrow, true))
-    for (int i=0; i<nrow; ++i) {
-      int val = xP[i];
-      if (val!=NA_INTEGER) continue;
-      int j=0; while (val==NA_INTEGER && j<k) val=((int *)valP[j++])[i];  // k is the number of non-singletons
-      if (val!=NA_INTEGER) xP[i]=val; else if (final) xP[i]=finalVal;
+
+    const bool final = finalVal != NA_INTEGER;
+
+#pragma omp parallel for num_threads(getDTthreads(nrow, true))
+    for (int i = 0; i < nrow; ++i) {
+      if (xP[i] != NA_INTEGER) continue;
+      int v = NA_INTEGER, j = 0;
+      while (v == NA_INTEGER && j < k)
+        v = ((int *)valP[j++])[i];
+      if (v != NA_INTEGER) xP[i] = v;
+      else if (final) xP[i] = finalVal;
     }
   } break;
+
   case REALSXP: {
-    if (INHERITS(first, char_integer64)) {
-      int64_t *xP=(int64_t *)REAL(first), finalVal=NA_INTEGER64;
-      int k=0;
-      for (int j=0; j<nval; ++j) {
-        SEXP item = VECTOR_ELT(x, j+off);
-        if (length(item)==1) {
-          int64_t tt = ((int64_t *)REAL(item))[0];
-          if (tt==NA_INTEGER64) continue;
-          finalVal = tt;
+    double *xP = REAL(first), finalVal = NA_REAL;
+    int k = 0;
+
+    for (int j = 0; j < nval; ++j) {
+      SEXP item = VECTOR_ELT(x, j + off);
+      if (length(item) == 1) {
+        double v = REAL(item)[0];
+        if (!ISNA(v) && (!nan_is_na || !ISNAN(v))) {
+          finalVal = v;
           break;
         }
-        valP[k++] = REAL_RO(item);
-      }
-      const bool final = (finalVal!=NA_INTEGER64);
-      #pragma omp parallel for num_threads(getDTthreads(nrow, true))
-      for (int i=0; i<nrow; ++i) {
-        int64_t val=xP[i];
-        if (val!=NA_INTEGER64) continue;
-        int j=0; while (val==NA_INTEGER64 && j<k) val=((int64_t *)valP[j++])[i];
-        if (val!=NA_INTEGER64) xP[i]=val; else if (final) xP[i]=finalVal;
-      }
-    } else {
-      double *xP = REAL(first), finalVal=NA_REAL;
-      int k=0;
-      if (nan_is_na) {
-        for (int j=0; j<nval; ++j) {
-          SEXP item = VECTOR_ELT(x, j+off);
-          if (length(item)==1) {
-            double tt = REAL(item)[0];
-            if (ISNAN(tt)) continue;
-            finalVal = tt;
-            break;
-          }
-          valP[k++] = REAL_RO(item);
-        }
-        const bool final = !ISNAN(finalVal);
-        #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;
-        }
-      } else {
-        for (int j=0; j<nval; ++j) {
-          SEXP item = VECTOR_ELT(x, j+off);
-          if (length(item)==1) {
-            double tt = REAL(item)[0];
-            if (ISNA(tt)) continue;
-            finalVal = tt;
-            break;
-          }
-          valP[k++] = REAL_RO(item);
-        }
-        const bool final = !ISNA(finalVal);
-        #pragma omp parallel for num_threads(getDTthreads(nrow, true))
-        for (int i=0; i<nrow; ++i) {
-          double val=xP[i];
-          if (!ISNA(val)) continue;
-          int j=0; while (ISNA(val) && j<k) val=((double *)valP[j++])[i];
-          if (!ISNA(val)) xP[i]=val; else if (final) xP[i]=finalVal;
-        }
       }
+      valP[k++] = REAL_RO(item);
     }
-  } break;
-  case CPLXSXP: {
-    Rcomplex *xP = COMPLEX(first), finalVal=NA_CPLX;
-    int k=0;
-    for (int j=0; j<nval; ++j) {
-      SEXP item = VECTOR_ELT(x, j+off);
-      if (length(item)==1) {
-        Rcomplex tt = COMPLEX(item)[0];
-        if (ISNAN(tt.r) && ISNAN(tt.i)) continue;
-        finalVal = tt;
-        break;
-      }
-      valP[k++] = COMPLEX_RO(item);
-    }
-    const bool final = !ISNAN(finalVal.r) && !ISNAN(finalVal.i);
-    #pragma omp parallel for num_threads(getDTthreads(nrow, true))
-    for (int i=0; i<nrow; ++i) {
-      Rcomplex val=xP[i];
-      if (!ISNAN(val.r) && !ISNAN(val.i)) continue;
-      int j=0; while (ISNAN(val.r) && ISNAN(val.i) && j<k) val=((Rcomplex *)valP[j++])[i];
-      if (!ISNAN(val.r) || !ISNAN(val.i)) xP[i]=val; else if (final) xP[i]=finalVal;
+
+    const bool final = !ISNA(finalVal) && (!nan_is_na || !ISNAN(finalVal));
+
+#pragma omp parallel for num_threads(getDTthreads(nrow, true))
+    for (int i = 0; i < nrow; ++i) {
+      double v = xP[i];
+      if (!ISNA(v) && (!nan_is_na || !ISNAN(v))) continue;
+      int j = 0;
+      while ((ISNA(v) || (nan_is_na && ISNAN(v))) && j < k)
+        v = ((double *)valP[j++])[i];
+      if (!ISNA(v) && (!nan_is_na || !ISNAN(v))) xP[i] = v;
+      else if (final) xP[i] = finalVal;
     }
   } break;
+
   case STRSXP: {
     const SEXP *xP = STRING_PTR_RO(first);
-    SEXP finalVal=NA_STRING;
-    int k=0;
-    for (int j=0; j<nval; ++j) {
-      SEXP item = VECTOR_ELT(x, j+off);
-      if (length(item)==1) {
-        SEXP tt = STRING_ELT(item,0);
-        if (tt==NA_STRING) continue;
-        finalVal = tt;
+    SEXP finalVal = NA_STRING;
+    int k = 0;
+
+    for (int j = 0; j < nval; ++j) {
+      SEXP item = VECTOR_ELT(x, j + off);
+      if (length(item) == 1 && STRING_ELT(item, 0) != NA_STRING) {
+        finalVal = STRING_ELT(item, 0);
         break;
       }
       valP[k++] = STRING_PTR_RO(item);
     }
-    const bool final = (finalVal!=NA_STRING);
-    for (int i=0; i<nrow; ++i) {
-      SEXP val = xP[i];
-      if (val!=NA_STRING) continue;
-      int j=0; while (val==NA_STRING && j<k) val=((SEXP *)valP[j++])[i];
-      if (val!=NA_STRING) SET_STRING_ELT(first, i, val); else if (final) SET_STRING_ELT(first, i, finalVal);
+
+    const bool final = finalVal != NA_STRING;
+
+    for (int i = 0; i < nrow; ++i) {
+      if (xP[i] != NA_STRING) continue;
+      SEXP v = NA_STRING;
+      int j = 0;
+      while (v == NA_STRING && j < k)
+        v = ((SEXP *)valP[j++])[i];
+      if (v != NA_STRING) SET_STRING_ELT(first, i, v);
+      else if (final) SET_STRING_ELT(first, i, finalVal);
     }
   } break;
+
   default:
-    error(_("Type '%s' is not supported"), type2char(TYPEOF(first))); // e.g. raw is tested
+    error(_("Type '%s' not supported."), type2char(TYPEOF(first)));
   }
+
   UNPROTECT(nprotect);
   return first;
-}
+}
\ No newline at end of file