Closes #1050. Added isReallyReal(), fixed another auto indexing issue.

README.md

@@ -69,6 +69,7 @@
       * While fixing `#932`, issues on subsetting `NA` were also spotted and fixed, for e.g., `DT[x==NA]` or `DT[!x==NA]`. 
       * Works fine when RHS is of `list` type - quite unusual operation but could happen. Closes [#961](https://github.com/Rdatatable/data.table/issues/961). Thanks to @Gsee for the minimal report.
       * Auto indexing errored in some cases when LHS and RHS were not of same type. This is fixed now. Closes [#957](https://github.com/Rdatatable/data.table/issues/957). Thanks to @GSee for the minimal report.
+      * `DT[x == 2.5]` where `x` is integer type resulted in `val` being coerced to integer (for binary search) and therefore returned incorrect result. This is now identified using the function `isReallyReal()` and if so, auto indexing is turned off. Closes [#1050](https://github.com/Rdatatable/data.table/issues/1050).
 
   7. `as.data.table.list` with list input having 0-length items, e.g. `x = list(a=integer(0), b=3:4)`. `as.data.table(x)` recycles item `a` with `NA`s to fit the length of the longer column `b` (length=2), as before now, but with an additional warning message that the item has been recycled with `NA`. Closes [#847](https://github.com/Rdatatable/data.table/issues/847). Thanks to @tvinodr for the report. This was a regression from 1.9.2.
 

inst/tests/tests.Rraw

@@ -6065,6 +6065,14 @@ test(1491.4, replace_dot(ee4), quote(c("a", "b") := list(.SD)))
 test(1491.5, replace_dot(ee5), quote(c("a", "b") := list(v1=x^2, v2 = list(.SD[[1L]]))))
 test(1491.6, replace_dot(ee6), quote(list(v1=list(.SD), v2=list(lm(. ~ xx)), v3=list(list(x)), v4=list(x^2))))
 
+# Fix for #1050
+dt = data.table(x=1:5, y=6:10)
+options(datatable.auto.index=FALSE)
+ans1 <- dt[x == 2.5]
+options(datatable.auto.index=TRUE)
+ans2 <- dt[x == 2.5]
+test(1492, ans1, ans2)
+
 ##########################
 
 

src/forder.c

@@ -1392,6 +1392,21 @@ SEXP isOrderedSubset(SEXP x, SEXP nrow)
 }
 */
 
+SEXP isReallyReal(SEXP x) {
+    int n, i=0;
+    SEXP ans;
+    if (!isReal(x))
+        error("x must be of type double.");
+    n = length(x);
+    ans = PROTECT(allocVector(LGLSXP, 1));
+    INTEGER(ans)[0] = FALSE;
+    while (i < n && REAL(x)[i] == abs(REAL(x)[i]))
+        i++;
+    if (i != n) LOGICAL(ans)[0] = TRUE;
+    UNPROTECT(1);
+    return(ans);
+}
+
 void pbin(unsigned long long n)
 // trace utility for dev to be used in gdb, since couldn't get stdlib::atoi() to link
 {

src/init.c

@@ -57,6 +57,7 @@ SEXP whichwrapper();
 SEXP shift();
 SEXP transpose();
 SEXP anyNA();
+SEXP isReallyReal();
 
 // .Externals
 SEXP fastmean();
@@ -116,6 +117,7 @@ R_CallMethodDef callMethods[] = {
 {"Cshift", (DL_FUNC) &shift, -1},
 {"Ctranspose", (DL_FUNC) &transpose, -1},
 {"CanyNA", (DL_FUNC) &anyNA, -1},
+{"CisReallyReal", (DL_FUNC) &isReallyReal, -1},
 {NULL, NULL, 0}
 };