[Improvement](JSONB) improve performance JSONB initial json parsing u…

…sing simdjson (apache#15219)

test data: https://data.gharchive.org/2020-11-13-18.json.gz, 2GB, 197696 lines
before: String 13s vs. JSONB 28s
after: String 13s vs. JSONB 16s

**NOTICE: simdjson need to be patched since BOOL is conflicted with a macro BOOL defined in odbc sqltypes.h**

.licenserc.yaml

@@ -52,6 +52,7 @@ header:
     - "be/src/util/jsonb_document.h"
     - "be/src/util/jsonb_error.h"
     - "be/src/util/jsonb_parser.h"
+    - "be/src/util/jsonb_parser_simd.h"
     - "be/src/util/jsonb_stream.h"
     - "be/src/util/jsonb_updater.h"
     - "be/src/util/jsonb_utils.h"

be/CMakeLists.txt

@@ -510,6 +510,8 @@ if ("${CMAKE_BUILD_TARGET_ARCH}" STREQUAL "x86" OR "${CMAKE_BUILD_TARGET_ARCH}"
     if (USE_AVX2)
         set(CXX_COMMON_FLAGS "${CXX_COMMON_FLAGS} -mavx2")
     endif()
+    # set -mlzcnt for leading zero count used by simdjson
+    set(CXX_COMMON_FLAGS "${CXX_COMMON_FLAGS} -msse4.2")
 endif()
 set(CXX_COMMON_FLAGS "${CXX_COMMON_FLAGS} -Wno-attributes -DS2_USE_GFLAGS -DS2_USE_GLOG")
 

be/src/runtime/jsonb_value.h

@@ -22,7 +22,7 @@
 #include "util/cpu_info.h"
 #include "util/hash_util.hpp"
 #include "util/jsonb_error.h"
-#include "util/jsonb_parser.h"
+#include "util/jsonb_parser_simd.h"
 #include "util/jsonb_utils.h"
 #include "vec/common/string_ref.h"
 
@@ -38,7 +38,7 @@ struct JsonBinaryValue {
     // default nullprt and size 0 for invalid or NULL value
     const char* ptr = nullptr;
     size_t len = 0;
-    JsonbParser parser;
+    JsonbParserSIMD parser;
 
     JsonBinaryValue() : ptr(nullptr), len(0) {}
     JsonBinaryValue(char* ptr, int len) { from_json_string(const_cast<const char*>(ptr), len); }

be/src/util/jsonb_error.h

@@ -30,12 +30,14 @@ enum class JsonbErrType {
     E_EMPTY_DOCUMENT,
     E_OUTPUT_FAIL,
     E_INVALID_DOCU,
+    E_INVALID_TYPE,
     E_INVALID_SCALAR_VALUE,
     E_INVALID_KEY_STRING,
     E_INVALID_KEY_LENGTH,
     E_INVALID_STR,
     E_INVALID_OBJ,
     E_INVALID_ARR,
+    E_INVALID_NUMBER,
     E_INVALID_HEX,
     E_INVALID_OCTAL,
     E_INVALID_DECIMAL,
@@ -53,6 +55,7 @@ enum class JsonbErrType {
     E_INVALID_JSONB_OBJ,
     E_NESTING_LVL_OVERFLOW,
     E_INVALID_DOCU_COMPAT,
+    E_EXCEPTION,
 
     // new error code should always be added above
     E_NUM_ERRORS
@@ -77,13 +80,15 @@ class JsonbErrMsg {
             "Invalid document version",
             "Empty document",
             "Fatal error in writing JSONB",
-            "Invalid document: document must be an object or an array",
+            "Invalid document",
+            "Invalid json value type",
             "Invalid scalar value",
             "Invalid key string",
             "Key length exceeds maximum size allowed (64 bytes)",
             "Invalid string value",
             "Invalid JSON object",
             "Invalid JSON array",
+            "Invalid number",
             "Invalid HEX number",
             "Invalid octal number",
             "Invalid decimal number",
@@ -100,7 +105,8 @@ class JsonbErrMsg {
             "Invalid update operation",
             "Invalid JSONB object (internal)",
             "Object or array has too many nesting levels",
-            "Invalid document: document must be an object or an array",
+            "Invalid document",
+            "Exception throwed",
 
             nullptr /* E_NUM_ERRORS */
     };