feat: Add Spark concat_ws function (#8854)

Summary:
Add concat_ws Spark function which returns the concatenation for the
input, separated by a separator (the first argument). It allows variable
number of VARCHAR or ARRAY\<VARCHAR\> arguments. And these two
types can be used in combination.

This function is a bit similar to [ConcatFunction](https://github.com/facebookincubator/velox/blob/main/velox/functions/prestosql/StringFunctions.cpp#L140), except that `concat_ws`
requires separator and supports using ARRAY<VARCHAR> type and mixed types.

This PR is based on #6292 (author: unigof). There are a few bug fixes
and improvements. Also made some changes to align with Spark.

Doc [link](https://docs.databricks.com/en/sql/language-manual/functions/concat_ws.html).

Pull Request resolved: #8854

Reviewed By: kgpai

Differential Revision: D66898251

Pulled By: bikramSingh91

fbshipit-source-id: 1fcd193a245bea4062c4e20d1e1db9ad6cc3290b

velox/docs/functions/spark/string.rst

@@ -25,6 +25,25 @@ String Functions
     If ``n < 0``, the result is an empty string.
     If ``n >= 256``, the result is equivalent to chr(``n % 256``).
 
+.. spark:function:: concat_ws(separator, [string/array<string>], ...) -> varchar
+
+   Returns the concatenation result for ``string`` and all elements in ``array<string>``, separated
+   by ``separator``. The first argument is ``separator`` whose type is VARCHAR. Then, this function
+   can take variable number of remaining arguments , and it allows mixed use of ``string`` type and
+   ``array<string>`` type. Skips NULL argument or NULL array element during the concatenation. If
+   ``separator`` is NULL, returns NULL, regardless of the following inputs. For non-NULL ``separator``,
+   if no remaining input exists or all remaining inputs are NULL, returns an empty string. ::
+
+        SELECT concat_ws('~', 'a', 'b', 'c'); -- 'a~b~c'
+        SELECT concat_ws('~', ['a', 'b', 'c'], ['d']); -- 'a~b~c~d'
+        SELECT concat_ws('~', 'a', ['b', 'c']); -- 'a~b~c'
+        SELECT concat_ws('~', '', [''], ['a', '']); -- '~~a~'
+        SELECT concat_ws(NULL, 'a'); -- NULL
+        SELECT concat_ws('~'); -- ''
+        SELECT concat_ws('~', NULL, [NULL], 'a', 'b'); -- 'a~b'
+        SELECT concat_ws('~', NULL, NULL); -- ''
+        SELECT concat_ws('~', [NULL]); -- ''
+
 .. spark:function:: contains(left, right) -> boolean
 
     Returns true if 'right' is found in 'left'. Otherwise, returns false. ::

velox/expression/fuzzer/ExpressionFuzzer.cpp

@@ -354,11 +354,23 @@ static void appendSpecialForms(
           },
           {
               "cast",
-              /// TODO: Add supported Cast signatures to CastTypedExpr and
-              /// expose
-              /// them to fuzzer instead of hard-coding signatures here.
+              // TODO: Add supported Cast signatures to CastTypedExpr and
+              // expose
+              // them to fuzzer instead of hard-coding signatures here.
               getSignaturesForCast(),
           },
+          {
+              // For Spark SQL only.
+              "concat_ws",
+              std::vector<facebook::velox::exec::FunctionSignaturePtr>{
+                  // Signature: concat_ws (separator, input, ...) -> output:
+                  // varchar, varchar, varchar, ... -> varchar
+                  facebook::velox::exec::FunctionSignatureBuilder()
+                      .argumentType("varchar")
+                      .variableArity("varchar")
+                      .returnType("varchar")
+                      .build()},
+          },
       };
 
   auto specialFormNames = splitNames(specialForms);

velox/functions/sparksql/CMakeLists.txt

@@ -18,6 +18,7 @@ velox_add_library(
   ArrayGetFunction.cpp
   ArraySort.cpp
   Comparisons.cpp
+  ConcatWs.cpp
   DecimalArithmetic.cpp
   DecimalCompare.cpp
   Hash.cpp