Fix split function

velox/functions/sparksql/SplitFunctions.cpp

@@ -18,20 +18,16 @@
 
 #include "velox/expression/VectorFunction.h"
 #include "velox/expression/VectorWriters.h"
+#include <boost/regex.hpp>
+#include <iostream>
 
 namespace facebook::velox::functions::sparksql {
 namespace {
 
-/// This class only implements the basic split version in which the pattern is a
-/// single character
 class SplitCharacter final : public exec::VectorFunction {
  public:
-  explicit SplitCharacter(const char pattern) : pattern_{pattern} {
-    static constexpr std::string_view kRegexChars = ".$|()[{^?*+\\";
-    VELOX_CHECK(
-        kRegexChars.find(pattern) == std::string::npos,
-        "This version of split supports single-length non-regex patterns");
-  }
+  explicit SplitCharacter(const std::string& pattern) : 
+    pattern_(boost::regex(pattern)) {}
 
   void apply(
       const SelectivityVector& rows,
@@ -50,8 +46,15 @@ class SplitCharacter final : public exec::VectorFunction {
       auto& arrayWriter = resultWriter.current();
 
       const StringView& current = input->valueAt<StringView>(row);
+      std::string str = current.str();
       const char* pos = current.begin();
       const char* end = pos + current.size();
+      boost::sregex_token_iterator posT(str.begin(), str.end(), pattern_, -1);
+      boost::sregex_token_iterator endT;
+      while (posT != endT) {
+        std::cout << "val: " <<  *posT++ << std::endl;;
+      }
+
       const char* delim;
       do {
         delim = std::find(pos, end, pattern_);
@@ -72,7 +75,7 @@ class SplitCharacter final : public exec::VectorFunction {
   }
 
  private:
-  const char pattern_;
+  boost::regex pattern_;
 };
 
 /// This class will be updated in the future as we support more variants of
@@ -91,12 +94,7 @@ class Split final : public exec::VectorFunction {
     VELOX_CHECK(
         delimiterVector, "Split function supports only constant delimiter");
     auto patternString = args[1]->as<ConstantVector<StringView>>()->valueAt(0);
-    VELOX_CHECK_EQ(
-        patternString.size(),
-        1,
-        "split only supports only single-character pattern");
-    char pattern = patternString.data()[0];
-    SplitCharacter splitCharacter(pattern);
+    SplitCharacter splitCharacter(patternString.str());
     splitCharacter.apply(rows, args, nullptr, context, result);
   }
 };
@@ -119,10 +117,9 @@ std::shared_ptr<exec::VectorFunction> createSplit(
   if (pattern.size() != 1) {
     return std::make_shared<Split>();
   }
-  char charPattern = pattern.data()[0];
   // TODO: Add support for zero-length pattern, 2-character pattern
   // TODO: add support for general regex pattern using R2
-  return std::make_shared<SplitCharacter>(charPattern);
+  return std::make_shared<SplitCharacter>(pattern.str());
 }
 
 std::vector<std::shared_ptr<exec::FunctionSignature>> signatures() {