mihaibudiu commented on code in PR #3689:
URL: https://github.com/apache/calcite/pull/3689#discussion_r1496769044


##########
testkit/src/main/java/org/apache/calcite/test/SqlOperatorTest.java:
##########
@@ -6222,6 +6222,49 @@ void checkRegexpExtract(SqlOperatorFixture f0, 
FunctionAlias functionAlias) {
     f.checkNull("log(10, cast(null as real))");
   }
 
+  /** Test case for
+   * <a 
href="https://issues.apache.org/jira/browse/CALCITE-6224";>[CALCITE-6224]
+   * Add LOG2 function (enabled in MYSQL, Spark library)</a>. */
+  @Test void testLog2Func() {
+    final SqlOperatorFixture f0 = fixture();
+    f0.checkFails("^log2(4)^",
+        "No match found for function signature LOG2\\(<NUMERIC>\\)", false);
+    final Consumer<SqlOperatorFixture> consumer = f -> {
+      f.setFor(SqlLibraryOperators.LOG2);
+      f.checkScalarApprox("log2(2)", "DOUBLE NOT NULL",
+          isWithin(1.0, 0.000001));
+      f.checkScalarApprox("log2(4)", "DOUBLE NOT NULL",
+          isWithin(2.0, 0.000001));
+      f.checkScalarApprox("log2(65536)", "DOUBLE NOT NULL",
+          isWithin(16.0, 0.000001));
+      f.checkScalarApprox("log2(2.0/3)", "DOUBLE NOT NULL",
+          isWithin(-0.5849625007211561, 0.000001));
+      f.checkScalarApprox("log2(4.0/3)", "DOUBLE NOT NULL",
+          isWithin(0.4150374992788435, 0.000001));
+      f.checkScalarApprox("log2(0.5)", "DOUBLE NOT NULL",
+          isWithin(-1.0, 0.000001));
+      f.checkScalarApprox("log2(cast(10e8 as double))", "DOUBLE NOT NULL",
+          isWithin(29.897352853986263, 0.000001));
+      f.checkScalarApprox("log2(cast(10e8 as float))", "DOUBLE NOT NULL",
+          isWithin(29.897352853986263, 0.000001));
+      f.checkScalarApprox("log2(1e+52)", "DOUBLE NOT NULL",

Review Comment:
   Most systems, including Calcite, use the IEEE 754 for floating point, which 
is a standard. MySQL apparently does not. So FLOAT/DOUBLE are really different 
types from the ones used in Java and Calcite. This means that it's probably 
impossible to emulate exactly the behavior of MySQL FP computations in Calcite. 
Even if you make LOG work, other cases won't, like the case when addition 
overflows. Calcite will happily use +Infinity while MySQL will stop with an 
error.



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscr...@calcite.apache.org

For queries about this service, please contact Infrastructure at:
us...@infra.apache.org

Reply via email to