Handle and cases

NVIDIA · Dec 4, 2021 · 57e9425 · 57e9425
1 parent ef4434c
commit 57e9425
Show file tree

Hide file tree

Showing 2 changed files with 27 additions and 15 deletions.
diff --git a/sql-plugin/src/main/scala/com/nvidia/spark/rapids/RegexParser.scala b/sql-plugin/src/main/scala/com/nvidia/spark/rapids/RegexParser.scala
@@ -169,11 +169,9 @@ class RegexParser(pattern: String) {
               throw new RegexUnsupportedException(
                 s"unexpected EOF while parsing escaped character", Some(pos))
             case Some(ch) =>
-              ch match {
-                case '\\' | '^' | '-' | ']' | '+' =>
-                  // escaped metacharacter within character class
-                  characterClass.appendEscaped(consumeExpected(ch))
-              }
+              // typically an escaped metacharacter ('\\' | '^' | '-' | ']' | '+')
+              // within the character class, but could be any escaped character
+              characterClass.appendEscaped(consumeExpected(ch))
           }
         case '\u0000' =>
           throw new RegexUnsupportedException(
@@ -508,23 +506,36 @@ class CudfRegexTranspiler(replace: Boolean) {
           // `[^a\r]`   => `(?:[\n]|[^a])`
           // `[^a\n]`   => `(?:[\r]|[^a])`
           // `[^a\r\n]` => `[^a]`
+          // `[^\r\n]`  => `[^\r\n]`
 
-          val newlineCharsInClass = characters.flatMap {
-            case RegexChar(ch) if ch == '\n' || ch == '\r' =>
-              Seq(ch)
-            case _ =>
-              Seq.empty
+          val allLinefeed = components.forall {
+            case RegexChar(ch) => ch == '\n' || ch == '\r'
+            case RegexEscaped(ch) => ch == 'n' || ch == 'r'
+            case _ => false
           }
+
+          val newlineCharsInClass = components.flatMap {
+            case RegexChar(ch) if ch == '\n' || ch == '\r' => Seq(ch)
+            case RegexEscaped(ch) if ch == 'n' => Seq('\n')
+            case RegexEscaped(ch) if ch == 'r' => Seq('\r')
+            case _ => Seq.empty
+          }.distinct
+
           val negatedNewlines = Seq('\r', '\n').diff(newlineCharsInClass)
-          if (negatedNewlines.isEmpty) {
-            RegexCharacterClass(negated, ListBuffer(components: _*))
+
+          if (allLinefeed && newlineCharsInClass.length == 2) {
+            // special case for `[^\r\n]`
+            RegexCharacterClass(negated = true, ListBuffer(components: _*))
+          } else if (negatedNewlines.isEmpty) {
+            RegexCharacterClass(negated = true, ListBuffer(components: _*))
           } else {
             RegexGroup(capture = false,
               RegexChoice(
                 RegexCharacterClass(negated = false,
                   characters = ListBuffer(negatedNewlines.map(RegexChar): _*)),
-                RegexCharacterClass(negated, ListBuffer(components: _*))))
+                RegexCharacterClass(negated = true, ListBuffer(components: _*))))
           }
+
         } else {
           RegexCharacterClass(negated, ListBuffer(components: _*))
         }

diff --git a/tests/src/test/scala/com/nvidia/spark/rapids/RegularExpressionTranspilerSuite.scala b/tests/src/test/scala/com/nvidia/spark/rapids/RegularExpressionTranspilerSuite.scala
@@ -249,8 +249,9 @@ class RegularExpressionTranspilerSuite extends FunSuite with Arm {
   }
 
   test("compare CPU and GPU: regexp replace negated character class") {
-    val inputs = Seq("a", "b", "a\nb")
-    val patterns = Seq("[^z]")
+    val inputs = Seq("a", "b", "a\nb", "a\r\nb\n\rc\rd")
+    val patterns = Seq("[^z]", "[^\r]", "[^\n]", "[^\r]", "[^\r\n]",
+      "[^a\n]", "[^b\r]", "[^bc\r\n]", "[^\\r\\n]")
     assertCpuGpuMatchesRegexpReplace(patterns, inputs)
   }