fix(core): support the Unicode literal for MSSQL

ibis-server/tests/routers/v3/connector/mssql/conftest.py

@@ -1,8 +1,13 @@
 import pathlib
 
+import pandas as pd
 import pytest
+import sqlalchemy
+from sqlalchemy import text
 from testcontainers.mssql import SqlServerContainer
 
+from tests.conftest import file_path
+
 pytestmark = pytest.mark.mssql
 
 base_url = "/v3/connector/mssql"
@@ -20,6 +25,21 @@ def mssql(request) -> SqlServerContainer:
     mssql = SqlServerContainer(
         "mcr.microsoft.com/mssql/server:2019-CU27-ubuntu-20.04", dialect="mssql+pyodbc"
     ).start()
+    engine = sqlalchemy.create_engine(
+        f"{mssql.get_connection_url()}?driver=ODBC+Driver+18+for+SQL+Server&TrustServerCertificate=YES"
+    )
+    pd.read_parquet(file_path("resource/tpch/data/orders.parquet")).to_sql(
+        "orders", engine, index=False
+    )
+    with engine.begin() as conn:
+        conn.execute(text("CREATE TABLE unicode_test (id INT, letter NVARCHAR(10))"))
+        conn.execute(
+            text("INSERT INTO unicode_test (id, letter) VALUES (1, N'真夜中')")
+        )
+        conn.execute(
+            text("INSERT INTO unicode_test (id, letter) VALUES (2, 'ZUTOMAYO')")
+        )
+
     request.addfinalizer(mssql.stop)
     return mssql
 

ibis-server/tests/routers/v3/connector/mssql/test_query.py

@@ -0,0 +1,65 @@
+import base64
+
+import orjson
+import pytest
+
+from app.dependencies import (
+    X_WREN_FALLBACK_DISABLE,
+)
+from tests.routers.v3.connector.mssql.conftest import base_url
+
+manifest = {
+    "catalog": "wren",
+    "schema": "public",
+    "models": [
+        {
+            "name": "unicode_test",
+            "tableReference": {
+                "schema": "dbo",
+                "table": "unicode_test",
+            },
+            "columns": [
+                {"name": "id", "type": "integer"},
+                {"name": "letter", "type": "varchar"},
+            ],
+        }
+    ],
+    "dataSource": "MSSQL",
+}
+
+
+@pytest.fixture(scope="module")
+def manifest_str():
+    return base64.b64encode(orjson.dumps(manifest)).decode("utf-8")
+
+
+async def test_unicode_literal(client, manifest_str, connection_info):
+    response = await client.post(
+        url=f"{base_url}/query",
+        json={
+            "connectionInfo": connection_info,
+            "manifestStr": manifest_str,
+            "sql": "SELECT id FROM wren.public.unicode_test WHERE letter = '真夜中'",
+        },
+        headers={
+            X_WREN_FALLBACK_DISABLE: "true",
+        },
+    )
+    assert response.status_code == 200
+    result = response.json()
+    assert result["data"] == [[1]]
+
+    response = await client.post(
+        url=f"{base_url}/query",
+        json={
+            "connectionInfo": connection_info,
+            "manifestStr": manifest_str,
+            "sql": "SELECT id FROM wren.public.unicode_test WHERE letter = 'ZUTOMAYO'",
+        },
+        headers={
+            X_WREN_FALLBACK_DISABLE: "true",
+        },
+    )
+    assert response.status_code == 200
+    result = response.json()
+    assert result["data"] == [[2]]