Skip to content

Open Source alternative to Algolia + Pinecone and an Easier-to-Use alternative to ElasticSearch ⚡ 🔍 ✨ Fast, typo tolerant, in-memory fuzzy Search Engine for building delightful search experiences

Support emplace action for ingestion.

Kishore Nallan3a8a2fcadbd7
diff --git a/include/index.h b/include/index.h
index ef85c1f5..7dbb3da3 100644
--- a/include/index.h
+++ b/include/index.h
@@ -332,6 +332,7 @@ enum index_operation_t {
CREATE,
UPSERT,
UPDATE,
+ EMPLACE,
DELETE
};
diff --git a/src/collection.cpp b/src/collection.cpp
index 1cd40e34..e93e1c2b 100644
--- a/src/collection.cpp
+++ b/src/collection.cpp
@@ -99,7 +99,7 @@ Option<doc_seq_id_t> Collection::to_doc(const std::string & json_str, nlohmann::
if(operation == UPDATE) {
return Option<doc_seq_id_t>(400, "For update, the `id` key must be provided.");
}
- // for UPSERT or CREATE, if a document does not have an ID, we will treat it as a new doc
+ // for UPSERT, EMPLACE or CREATE, if a document does not have an ID, we will treat it as a new doc
uint32_t seq_id = get_next_seq_id();
document["id"] = std::to_string(seq_id);
return Option<doc_seq_id_t>(doc_seq_id_t{seq_id, true});
@@ -123,7 +123,7 @@ Option<doc_seq_id_t> Collection::to_doc(const std::string & json_str, nlohmann::
return Option<doc_seq_id_t>(409, std::string("A document with id ") + doc_id + " already exists.");
}
- // UPSERT or UPDATE
+ // UPSERT, EMPLACE or UPDATE
uint32_t seq_id = (uint32_t) std::stoul(seq_id_str);
return Option<doc_seq_id_t>(doc_seq_id_t{seq_id, false});
@@ -132,7 +132,7 @@ Option<doc_seq_id_t> Collection::to_doc(const std::string & json_str, nlohmann::
// for UPDATE, a document with given ID must be found
return Option<doc_seq_id_t>(404, "Could not find a document with id: " + doc_id);
} else {
- // for UPSERT or CREATE, if a document with given ID is not found, we will treat it as a new doc
+ // for UPSERT, EMPLACE or CREATE, if a document with given ID is not found, we will treat it as a new doc
uint32_t seq_id = get_next_seq_id();
return Option<doc_seq_id_t>(doc_seq_id_t{seq_id, true});
}
diff --git a/src/core_api.cpp b/src/core_api.cpp
index 5dfd7d8c..10f8b6ff 100644
--- a/src/core_api.cpp
+++ b/src/core_api.cpp
@@ -97,6 +97,8 @@ index_operation_t get_index_operation(const std::string& action) {
return UPDATE;
} else if(action == "upsert") {
return UPSERT;
+ } else if(action == "emplace") {
+ return EMPLACE;
}
return CREATE;
@@ -586,7 +588,8 @@ bool post_import_documents(const std::shared_ptr<http_req>& req, const std::shar
return false;
}
- if(req->params[ACTION] != "create" && req->params[ACTION] != "update" && req->params[ACTION] != "upsert") {
+ if(req->params[ACTION] != "create" && req->params[ACTION] != "update" && req->params[ACTION] != "upsert" &&
+ req->params[ACTION] != "emplace") {
res->final = true;
res->set_400("Parameter `" + std::string(ACTION) + "` must be a create|update|upsert.");
stream_response(req, res);
@@ -703,7 +706,8 @@ bool post_add_document(const std::shared_ptr<http_req>& req, const std::shared_p
req->params[ACTION] = "create";
}
- if(req->params[ACTION] != "create" && req->params[ACTION] != "update" && req->params[ACTION] != "upsert") {
+ if(req->params[ACTION] != "create" && req->params[ACTION] != "update" && req->params[ACTION] != "upsert" &&
+ req->params[ACTION] != "emplace") {
res->set_400("Parameter `" + std::string(ACTION) + "` must be a create|update|upsert.");
return false;
}
diff --git a/test/collection_test.cpp b/test/collection_test.cpp
index 09229ed0..159ecec9 100644
--- a/test/collection_test.cpp
+++ b/test/collection_test.cpp
@@ -1246,6 +1246,109 @@ TEST_F(CollectionTest, ImportDocumentsUpsert) {
ASSERT_EQ(409, import_results[1]["code"].get<size_t>());
}
+TEST_F(CollectionTest, ImportDocumentsEmplace) {
+ Collection* coll1;
+ std::vector<field> fields = {
+ field("title", field_types::STRING, false, false),
+ field("points", field_types::INT32, false, true)
+ };
+
+ coll1 = collectionManager.get_collection("coll1").get();
+ if (coll1 == nullptr) {
+ coll1 = collectionManager.create_collection("coll1", 4, fields).get();
+ }
+
+ nlohmann::json document;
+ std::vector<std::string> records = {R"({"id": "0", "title": "The Matrix", "points":0})",
+ R"({"id": "1", "title": "Inception", "points":1})"};
+
+ // use `emplace` mode for creating documents
+ auto import_response = coll1->add_many(records, document, EMPLACE);
+
+ ASSERT_TRUE(import_response["success"].get<bool>());
+ ASSERT_EQ(2, import_response["num_imported"].get<int>());
+
+ std::vector<nlohmann::json> import_results = import_res_to_json(records);
+ ASSERT_EQ(2, import_results.size());
+
+ for (size_t i = 0; i < 2; i++) {
+ ASSERT_TRUE(import_results[i]["success"].get<bool>());
+ ASSERT_EQ(1, import_results[i].size());
+ }
+
+ auto res = coll1->search("*", {}, "", {}, {}, {0}, 10, 1, token_ordering::FREQUENCY, {true}, 10).get();
+ ASSERT_EQ(2, res["found"].get<size_t>());
+
+ // emplace both update + create
+ records = {R"({"id": "1", "title": "The Inception"})",
+ R"({"id": "2", "title": "Spiderman", "points":2})"};
+
+ import_response = coll1->add_many(records, document, EMPLACE);
+
+ ASSERT_TRUE(import_response["success"].get<bool>());
+ ASSERT_EQ(2, import_response["num_imported"].get<int>());
+
+ import_results = import_res_to_json(records);
+ ASSERT_EQ(2, import_results.size());
+
+ for (size_t i = 0; i < 2; i++) {
+ ASSERT_TRUE(import_results[i]["success"].get<bool>());
+ ASSERT_EQ(1, import_results[i].size());
+ }
+
+ res = coll1->search("*", {}, "", {}, {}, {0}, 10, 1, token_ordering::FREQUENCY, {true}, 10).get();
+ ASSERT_EQ(3, res["found"].get<size_t>());
+
+ ASSERT_EQ("2", res["hits"][0]["document"]["id"].get<std::string>());
+ ASSERT_EQ(2, res["hits"][0]["document"]["points"].get<size_t>());
+
+ ASSERT_EQ("1", res["hits"][1]["document"]["id"].get<std::string>());
+ ASSERT_EQ(1, res["hits"][1]["document"]["points"].get<size_t>());
+ ASSERT_EQ("The Inception", res["hits"][1]["document"]["title"].get<std::string>());
+
+ ASSERT_EQ("0", res["hits"][2]["document"]["id"].get<std::string>());
+ ASSERT_EQ(0, res["hits"][2]["document"]["points"].get<size_t>());
+
+ // emplace with an error due to bad data
+ records = {R"({"id": "2", "points": "abcd"})",
+ R"({"id": "3", "title": "Superman", "points":3})"};
+
+ import_response = coll1->add_many(records, document, EMPLACE);
+
+ ASSERT_FALSE(import_response["success"].get<bool>());
+ ASSERT_EQ(1, import_response["num_imported"].get<int>());
+
+ import_results = import_res_to_json(records);
+
+ ASSERT_EQ(2, import_results.size());
+
+ ASSERT_FALSE(import_results[0]["success"].get<bool>());
+
+ ASSERT_TRUE(import_results[1]["success"].get<bool>());
+ ASSERT_EQ(1, import_results[1].size());
+ ASSERT_EQ(1, import_results[1].size());
+
+ // can update individual document via "emplace"
+ std::string doc_3_update = R"({"id": "3", "title": "The Superman"})";
+ auto add_op = coll1->add(doc_3_update, EMPLACE);
+ ASSERT_TRUE(add_op.ok());
+
+ res = coll1->search("superman", {"title"}, "", {}, {}, {0}, 10, 1, token_ordering::FREQUENCY, {true}, 10).get();
+ ASSERT_EQ(1, res["found"].get<size_t>());
+
+ ASSERT_EQ("3", res["hits"][0]["document"]["id"].get<std::string>());
+ ASSERT_EQ(3, res["hits"][0]["document"]["points"].get<size_t>());
+ ASSERT_EQ("The Superman", res["hits"][0]["document"]["title"].get<std::string>());
+
+ // can create individual document via "emplace"
+ std::string doc_4_create = R"({"id": "4", "title": "The Avengers", "points": 4})";
+ add_op = coll1->add(doc_4_create, EMPLACE);
+ ASSERT_TRUE(add_op.ok());
+
+ res = coll1->search("*", {}, "", {}, {}, {0}, 10, 1, token_ordering::FREQUENCY, {true}, 10).get();
+ ASSERT_EQ(5, res["found"].get<size_t>());
+}
+
TEST_F(CollectionTest, DISABLED_CrashTroubleshooting) {
Collection *coll1;
std::vector<field> fields = {