add extract (#4098)

* add extract * format * fix typesignature * add a test case with no-ascii string * add test case in tck * fix clang compile error Co-authored-by: Sophie <84560950+Sophie-Xie@users.noreply.github.com> Co-authored-by: Yee <2520865+yixinglu@users.noreply.github.com> confirm with Yee
vesoft-inc · Apr 11, 2022 · 63d162c · 63d162c
1 parent 0fbe9f5
commit 63d162c
Show file tree

Hide file tree

Showing 3 changed files with 53 additions and 0 deletions.
diff --git a/src/common/function/FunctionManager.cpp b/src/common/function/FunctionManager.cpp
@@ -417,6 +417,7 @@ std::unordered_map<std::string, std::vector<TypeSignature>> FunctionManager::typ
     {"duration",
      {TypeSignature({Value::Type::STRING}, Value::Type::DURATION),
       TypeSignature({Value::Type::MAP}, Value::Type::DURATION)}},
+    {"extract", {TypeSignature({Value::Type::STRING, Value::Type::STRING}, Value::Type::LIST)}},
 };
 
 // static
@@ -2705,6 +2706,25 @@ FunctionManager::FunctionManager() {
       }
     };
   }
+  {
+    auto &attr = functions_["extract"];
+    attr.minArity_ = 2;
+    attr.maxArity_ = 2;
+    attr.isPure_ = true;
+    attr.body_ = [](const auto &args) -> Value {
+      if (!args[0].get().isStr() || !args[1].get().isStr()) {
+        return Value::kNullBadType;
+      }
+
+      const auto &s = args[0].get().getStr();
+      std::regex rgx(args[1].get().getStr());
+      List res;
+      for (std::sregex_iterator beg(s.begin(), s.end(), rgx), end{}; beg != end; ++beg) {
+        res.emplace_back(beg->str());
+      }
+      return res;
+    };
+  }
 }  // NOLINT
 
 // static

diff --git a/src/common/function/test/FunctionManagerTest.cpp b/src/common/function/test/FunctionManagerTest.cpp
@@ -341,6 +341,19 @@ TEST_F(FunctionManagerTest, functionCall) {
                   std::vector<Value>({".", 1, true, Value::kNullValue, "world", time}),
                   "1.true.world.09:39:21.000012000");
   }
+  {
+    TEST_FUNCTION(extract,
+                  std::vector<Value>({"Hello Nebula Graph", "\\w+"}),
+                  List({"Hello", "Nebula", "Graph"}));
+    TEST_FUNCTION(extract,
+                  std::vector<Value>({"你好，Nebula 图数据库", "(图|数据库|测试|Nebula)"}),
+                  List({"Nebula", "图", "数据库"}));
+    TEST_FUNCTION(extract,
+                  std::vector<Value>({"Hello Nebula Graph", "(Nebula|Graph)"}),
+                  List({"Nebula", "Graph"}));
+    TEST_FUNCTION(extract, std::vector<Value>({"Hello Nebula Graph", "testing"}), List());
+    TEST_FUNCTION(extract, std::vector<Value>({"Hello 2022 year", "\\d+"}), List({"2022"}));
+  }
   {
     TEST_FUNCTION(toBoolean, args_["int"], Value::kNullBadType);
     TEST_FUNCTION(toBoolean, args_["float"], Value::kNullBadType);

diff --git a/tests/tck/features/expression/FunctionCall.feature b/tests/tck/features/expression/FunctionCall.feature
@@ -121,6 +121,26 @@ Feature: Function Call Expression
       | result |
       | NULL   |
 
+  Scenario: extract
+    When executing query:
+      """
+      MATCH (a:player)-[b:serve]-(c:team{name: "Lakers"})
+      WHERE a.player.age > 45
+      RETURN extract(a.player.name, "\\w+") as result
+      """
+    Then the result should be, in any order:
+      | result                     |
+      | ["Shaquille", "O", "Neal"] |
+    When executing query:
+      """
+      MATCH (a:player)-[b:serve]-(c:team{name: "Lakers"})
+      WHERE a.player.age > 45
+      RETURN extract(a.player.name, "hello") as result
+      """
+    Then the result should be, in any order:
+      | result |
+      | []     |
+
   Scenario: round
     When executing query:
       """