Use parse_url kernel for PROTOCOL parsing

docs/additional-functionality/advanced_configs.md

@@ -300,6 +300,7 @@ Name | SQL Function(s) | Description | Default Value | Notes
 <a name="sql.expression.NthValue"></a>spark.rapids.sql.expression.NthValue|`nth_value`|nth window operator|true|None|
 <a name="sql.expression.OctetLength"></a>spark.rapids.sql.expression.OctetLength|`octet_length`|The byte length of string data|true|None|
 <a name="sql.expression.Or"></a>spark.rapids.sql.expression.Or|`or`|Logical OR|true|None|
+<a name="sql.expression.ParseUrl"></a>spark.rapids.sql.expression.ParseUrl|`parse_url`|Extracts a part from a URL|true|None|
 <a name="sql.expression.PercentRank"></a>spark.rapids.sql.expression.PercentRank|`percent_rank`|Window function that returns the percent rank value within the aggregation window|true|None|
 <a name="sql.expression.Pmod"></a>spark.rapids.sql.expression.Pmod|`pmod`|Pmod|true|None|
 <a name="sql.expression.PosExplode"></a>spark.rapids.sql.expression.PosExplode|`posexplode_outer`, `posexplode`|Given an input array produces a sequence of rows for each value in the array|true|None|

docs/compatibility.md

@@ -451,6 +451,17 @@ Spark stores timestamps internally relative to the JVM time zone.  Converting an
 between time zones is not currently supported on the GPU. Therefore operations involving timestamps
 will only be GPU-accelerated if the time zone used by the JVM is UTC.
 
+## URL parsing
+
+`parse_url` can produce different results on the GPU compared to the CPU. 
+
+Known issues for PROTOCOL parsing:
+- If urls containing utf-8 special characters, PROTOCOL results on GPU will be null.
+- If urls containing ipv6 host, GPU will return null for PROTOCOL.
+- GPU will still try to parse the PROTOCOL instead of returning null for some edge invalid cases,
+  such as urls containing multiple '#' in REF (http://##) or empty authority component followed by 
+  a empty path (http://).
+
 ## Windowing
 
 ### Window Functions