Add integration test for RLike with embedded null in input

docs/compatibility.md

@@ -594,13 +594,10 @@ Spark uses `^` and `$` to refer to the start and end of the entire string (equiv
 
 As a workaround, `\A` and `\Z` can be used instead of `^` and `$`.
 
-### Null character in input
+### Null support
 
-The GPU implementation of RLike will not match anything after a null character within a string.
-
-| Pattern   | Input     | Spark on CPU | Spark on GPU |
-|-----------|-----------|--------------|--------------|
-| `A`       | `\u0000A` | Match        | No Match     |
+The GPU implementation of RLike supports null characters in the input but does not support null characters in 
+the regular expression and will fall back to the CPU in this case.
 
 ### Qualifiers with nothing to repeat
 

integration_tests/src/main/python/string_test.py

@@ -14,7 +14,7 @@
 
 import pytest
 
-from asserts import assert_gpu_and_cpu_are_equal_collect, assert_gpu_and_cpu_are_equal_sql, assert_gpu_sql_fallback_collect
+from asserts import assert_gpu_and_cpu_are_equal_collect, assert_gpu_and_cpu_are_equal_sql, assert_gpu_sql_fallback_collect, assert_gpu_fallback_collect
 from conftest import is_databricks_runtime
 from data_gen import *
 from marks import *
@@ -461,6 +461,26 @@ def test_rlike():
                 'a rlike "a[bc]d"'),
             conf={'spark.rapids.sql.expression.RLike': 'true'})
 
+def test_rlike_embedded_null():
+    gen = mk_str_gen('[abcd]{1,3}')\
+            .with_special_case('\u0000aaa')
+    assert_gpu_and_cpu_are_equal_collect(
+            lambda spark: unary_op_df(spark, gen).selectExpr(
+                'a rlike "a{2}"',
+                'a rlike "a{1,3}"',
+                'a rlike "a{1,}"',
+                'a rlike "a[bc]d"'),
+            conf={'spark.rapids.sql.expression.RLike': 'true'})
+
+@allow_non_gpu('ProjectExec', 'RLike')
+def test_rlike_null_pattern():
+    gen = mk_str_gen('[abcd]{1,3}')
+    assert_gpu_fallback_collect(
+            lambda spark: unary_op_df(spark, gen).selectExpr(
+                'a rlike "a\u0000"'),
+            'RLike',
+            conf={'spark.rapids.sql.expression.RLike': 'true'})
+
 def test_rlike_escape():
     gen = mk_str_gen('[ab]{0,2}[\\-\\+]{0,2}')
     assert_gpu_and_cpu_are_equal_collect(

sql-plugin/src/main/scala/org/apache/spark/sql/rapids/stringFunctions.scala

@@ -751,8 +751,15 @@ class GpuRLikeMeta(
     rule: DataFromReplacementRule) extends BinaryExprMeta[RLike](expr, conf, parent, rule) {
 
     override def tagExprForGpu(): Unit = {
-      if (!expr.right.isInstanceOf[Literal]) {
-        willNotWorkOnGpu(s"RLike with non-literal pattern is not supported on GPU")
+      expr.right match {
+        case Literal(str: UTF8String, _) =>
+          if (str.toString.contains("\u0000")) {
+            // see https://github.com/NVIDIA/spark-rapids/issues/3962
+            willNotWorkOnGpu("The GPU implementation of RLike does not " +
+              "support null characters in the pattern")
+          }
+        case _ =>
+          willNotWorkOnGpu(s"RLike with non-literal pattern is not supported on GPU")
       }
     }