Fall back to CPU for regular expressions containing hex digits (#4492)

* Fall back to CPU for regular expressions containing hex digits

Signed-off-by: Andy Grove <andygrove@nvidia.com>

* assert that hex and octal digits fall back to CPU

Signed-off-by: Andy Grove <andygrove@nvidia.com>

* add newline between tests

docs/compatibility.md

@@ -475,6 +475,7 @@ Here are some examples of regular expression patterns that are not supported on
 - Beginning-of-line and end-of-line anchors (`^` and `$`) are not supported in some contexts, such as when combined 
   with a choice (`^|a`).
 - String anchors `\z` and `\Z` are not supported by `regexp_replace`
+- Hex and octal digits
 
 In addition to these cases that can be detected, there are also known issues that can cause incorrect results:
 

sql-plugin/src/main/scala/com/nvidia/spark/rapids/RegexParser.scala

@@ -456,13 +456,14 @@ class CudfRegexTranspiler(replace: Boolean) {
       }
 
       case RegexOctalChar(_) =>
-        // cuDF produced different results compared to Spark in some cases
-        // example: "a\141|.$"
+        // see https://github.com/NVIDIA/spark-rapids/issues/4288
         throw new RegexUnsupportedException(
           s"cuDF does not support octal digits consistently with Spark")
 
       case RegexHexDigit(_) =>
-        regex
+        // see https://github.com/NVIDIA/spark-rapids/issues/4486
+        throw new RegexUnsupportedException(
+          s"cuDF does not support hex digits consistently with Spark")
 
       case RegexEscaped(ch) => ch match {
         case 'b' | 'B' =>

tests/src/test/scala/com/nvidia/spark/rapids/RegularExpressionTranspilerSuite.scala

@@ -131,6 +131,22 @@ class RegularExpressionTranspilerSuite extends FunSuite with Arm {
         "cuDF does not support null characters in regular expressions"))
   }
 
+  test("cuDF does not support hex digits consistently with Spark") {
+    // see https://github.com/NVIDIA/spark-rapids/issues/4486
+    val patterns = Seq(raw"\xA9", raw"\x00A9", raw"\x10FFFF")
+    patterns.foreach(pattern =>
+      assertUnsupported(pattern, replace = false,
+        "cuDF does not support hex digits consistently with Spark"))
+  }
+
+  test("cuDF does not support octal digits consistently with Spark") {
+    // see https://github.com/NVIDIA/spark-rapids/issues/4288
+    val patterns = Seq(raw"\07", raw"\077", raw"\0377")
+    patterns.foreach(pattern =>
+      assertUnsupported(pattern, replace = false,
+        "cuDF does not support octal digits consistently with Spark"))
+  }
+
   test("string anchors - find") {
     val patterns = Seq("\\Atest", "test\\z", "test\\Z")
     assertCpuGpuMatchesRegexpFind(patterns, Seq("", "test", "atest", "testa",
@@ -159,11 +175,6 @@ class RegularExpressionTranspilerSuite extends FunSuite with Arm {
     assertCpuGpuMatchesRegexpFind(Seq(pattern), Seq("1\r2", "1\n2", "1\r\n2"))
   }
 
-  ignore("known issue - octal digit") {
-    val pattern = "a\\141|.$" // using hex works fine e.g. "a\\x61|.$"
-    assertCpuGpuMatchesRegexpFind(Seq(pattern), Seq("] b["))
-  }
-
   test("character class with ranges") {
     val patterns = Seq("[a-b]", "[a-zA-Z]")
     patterns.foreach(parse)
@@ -240,12 +251,6 @@ class RegularExpressionTranspilerSuite extends FunSuite with Arm {
     assertCpuGpuMatchesRegexpFind(patterns, inputs)
   }
 
-  test("compare CPU and GPU: hex") {
-    val patterns = Seq(raw"\x61")
-    val inputs = Seq("a", "b")
-    assertCpuGpuMatchesRegexpFind(patterns, inputs)
-  }
-
   test("compare CPU and GPU: octal") {
     val patterns = Seq("\\\\141")
     val inputs = Seq("a", "b")