[improve](json)improve json support empty keys (apache#36762)

support empty key for json behavior reference MSYQL
eldenmoon · Jul 1, 2024 · 6a9f003 · 6a9f003
1 parent 707cb40
commit 6a9f003
Show file tree

Hide file tree

Showing 11 changed files with 3,739 additions and 5 deletions.
diff --git a/be/src/util/jsonb_document.h b/be/src/util/jsonb_document.h
@@ -453,6 +453,7 @@ typedef std::underlying_type<JsonbType>::type JsonbTypeUnder;
  */
 class JsonbKeyValue {
 public:
+    // now we use sMaxKeyId to represent an empty key
     static const int sMaxKeyId = 65535;
     typedef uint16_t keyid_type;
 

diff --git a/be/src/util/jsonb_utils.h b/be/src/util/jsonb_utils.h
@@ -191,7 +191,12 @@ class JsonbToJson {
             if (iter->klen()) {
                 string_to_json(iter->getKeyStr(), iter->klen());
             } else {
-                os_.write(iter->getKeyId());
+                // NOTE: we use sMaxKeyId to represent an empty key. see jsonb_writer.h
+                if (iter->getKeyId() == JsonbKeyValue::sMaxKeyId) {
+                    string_to_json(nullptr, 0);
+                } else {
+                    os_.write(iter->getKeyId());
+                }
             }
             os_.put(':');
 

diff --git a/be/src/util/jsonb_writer.h b/be/src/util/jsonb_writer.h
@@ -79,7 +79,7 @@ class JsonbWriterT {
 
     // write a key string (or key id if an external dict is provided)
     uint32_t writeKey(const char* key, uint8_t len, hDictInsert handler = nullptr) {
-        if (len && !stack_.empty() && verifyKeyState()) {
+        if (!stack_.empty() && verifyKeyState()) {
             int key_id = -1;
             if (handler) {
                 key_id = handler(key, len);
@@ -88,9 +88,16 @@ class JsonbWriterT {
             uint32_t size = sizeof(uint8_t);
             if (key_id < 0) {
                 os_->put(len);
-                os_->write(key, len);
-                size += len;
-            } else if (key_id <= JsonbKeyValue::sMaxKeyId) {
+                if (len == 0) {
+                    // NOTE: we use sMaxKeyId to represent an empty key
+                    JsonbKeyValue::keyid_type idx = JsonbKeyValue::sMaxKeyId;
+                    os_->write((char*)&idx, sizeof(JsonbKeyValue::keyid_type));
+                    size += sizeof(JsonbKeyValue::keyid_type);
+                } else {
+                    os_->write(key, len);
+                    size += len;
+                }
+            } else if (key_id < JsonbKeyValue::sMaxKeyId) {
                 JsonbKeyValue::keyid_type idx = key_id;
                 os_->put(0);
                 os_->write((char*)&idx, sizeof(JsonbKeyValue::keyid_type));