Try using pymssql if the ODBC driver is unavailable

.github/workflows/ci.yml

@@ -34,14 +34,17 @@ jobs:
           python-version: ${{ matrix.python-version }}
 
       - name: Build the stack
-        run: docker-compose up -d mysql postgres presto trino clickhouse vertica
+        run: docker-compose up -d mysql postgres presto trino clickhouse vertica mssql
 
       - name: Install Poetry
         run: pip install poetry
 
       - name: Install package
         run: "poetry install"
 
+      - name: "Install MSSQL SDK for Python"
+        run: poetry install -E mssql
+
       # BigQuery start
     #   - id: 'auth'
     #     uses: 'google-github-actions/auth@v1'
@@ -68,6 +71,7 @@ jobs:
             DATADIFF_CLICKHOUSE_URI: 'clickhouse://clickhouse:Password1@localhost:9000/clickhouse'
             DATADIFF_VERTICA_URI: 'vertica://vertica:Password1@localhost:5433/vertica'
             DATADIFF_REDSHIFT_URI: '${{ secrets.DATADIFF_REDSHIFT_URI }}'
+            DATADIFF_MSSQL_URI: 'mssql://SA:Password123mssql@localhost:1433/master/dbo'
             MOTHERDUCK_TOKEN: '${{ secrets.MOTHERDUCK_TOKEN }}'
         run: |
           chmod +x tests/waiting_for_stack_up.sh

.github/workflows/ci_full.yml

@@ -34,14 +34,17 @@ jobs:
           python-version: ${{ matrix.python-version }}
 
       - name: Build the stack
-        run: docker-compose up -d mysql postgres presto trino clickhouse vertica
+        run: docker-compose up -d mysql postgres presto trino clickhouse vertica mssql
 
       - name: Install Poetry
         run: pip install poetry
 
       - name: Install package
         run: "poetry install"
 
+      - name: "Install MSSQL SDK for Python"
+        run: poetry install -E mssql
+
       # BigQuery start
     #   - id: 'auth'
     #     uses: 'google-github-actions/auth@v1'
@@ -64,6 +67,7 @@ jobs:
             DATADIFF_VERTICA_URI: 'vertica://vertica:Password1@localhost:5433/vertica'
             # DATADIFF_BIGQUERY_URI: '${{ secrets.DATADIFF_BIGQUERY_URI }}'
             DATADIFF_REDSHIFT_URI: '${{ secrets.DATADIFF_REDSHIFT_URI }}'
+            DATADIFF_MSSQL_URI: 'mssql://SA:Password123mssql@localhost:1433/master/dbo'
             MOTHERDUCK_TOKEN: '${{ secrets.MOTHERDUCK_TOKEN }}'
         run: |
           chmod +x tests/waiting_for_stack_up.sh

data_diff/databases/mssql.py

@@ -29,11 +29,27 @@
 )
 
 
+def _check_odbc_driver_installed():
+    import pyodbc
+
+    if "ODBC Driver 18 for SQL Server" in pyodbc.drivers():
+        return True
+    return False
+
+
+odbc_driver_installed = _check_odbc_driver_installed()
+
+
 @import_helper("mssql")
 def import_mssql():
-    import pyodbc
+    if odbc_driver_installed:
+        import pyodbc
 
-    return pyodbc
+        return pyodbc
+    else:
+        import pymssql
+
+        return pymssql
 
 
 @attrs.define(frozen=False)
@@ -170,23 +186,30 @@ class MsSQL(ThreadedDatabase):
     default_database: str
     _args: Dict[str, Any]
     _mssql: Any
+    _autocommit: bool
 
     def __init__(self, host, port, user, password, *, database, thread_count, **kw) -> None:
         super().__init__(thread_count=thread_count)
 
         args = dict(server=host, port=port, database=database, user=user, password=password, **kw)
         self._args = {k: v for k, v in args.items() if v is not None}
-        self._args["driver"] = "{ODBC Driver 18 for SQL Server}"
-
-        # TODO temp dev debug
-        self._args["TrustServerCertificate"] = "yes"
+        self._autocommit = False
 
         try:
             self.default_database = self._args["database"]
             self.default_schema = self._args["schema"]
         except KeyError:
             raise ValueError("Specify a default database and schema.")
 
+        if odbc_driver_installed:
+            self._args["driver"] = "{ODBC Driver 18 for SQL Server}"
+
+            # TODO temp dev debug
+            self._args["TrustServerCertificate"] = "yes"
+        else:
+            self._autocommit = True
+            self._args.pop("schema")
+
         self._mssql = None
 
     def create_connection(self):
@@ -227,3 +250,7 @@ def _query_cursor(self, c, sql_code: str):
             return super()._query_cursor(c, sql_code)
         except self._mssql.DatabaseError as e:
             raise QueryError(e)
+
+    @property
+    def is_autocommit(self) -> bool:
+        return self._autocommit

dev/dev.env

@@ -21,3 +21,7 @@ VERTICA_DB_NAME=vertica
 # leave VMART_DIR and VMART_ETL_SCRIPT empty.
 VMART_DIR=
 VMART_ETL_SCRIPT=
+
+# MSSQL credentials
+ACCEPT_EULA=Y
+MSSQL_SA_PASSWORD=Password123mssql

docker-compose.yml

@@ -117,13 +117,27 @@ services:
       networks:
       - local
 
+    mssql:
+      container_name: dd-mssql
+      image: mcr.microsoft.com/mssql/server
+      restart: always
+      volumes:
+        - mssql-data:/var/opt/mssql
+      ports:
+      - '1433:1433'
+      env_file:
+        - dev/dev.env
+      tty: true
+      networks:
+      - local
 
 
 volumes:
   postgresql-data:
   mysql-data:
   clickhouse-data:
   vertica-data:
+  mssql-data:
 
 networks:
   local:

pyproject.toml

@@ -43,6 +43,7 @@ vertica-python = {version="*", optional=true}
 urllib3 = "<2"
 oracledb = {version = "*", optional=true}
 pyodbc = {version=">=4.0.39", optional=true}
+pymssql = {version=">=2.2.11", optional=true}
 typing-extensions = ">=4.0.1"
 attrs = ">=23.1.0"
 mashumaro = {version = ">=2.9,<3.11.0", extras = ["msgpack"]}
@@ -74,15 +75,15 @@ redshift = ["psycopg2"]
 snowflake = ["snowflake-connector-python", "cryptography"]
 presto = ["presto-python-client"]
 oracle = ["oracledb"]
-mssql = ["pyodbc"]
+mssql = ["pyodbc", "pymssql"]
 # databricks = ["databricks-sql-connector"]
 trino = ["trino"]
 clickhouse = ["clickhouse-driver"]
 vertica = ["vertica-python"]
 duckdb = ["duckdb"]
 all-dbs = [
     "preql", "mysql-connector-python", "psycopg2", "snowflake-connector-python", "cryptography", "presto-python-client",
-    "oracledb", "pyodbc", "trino", "clickhouse-driver", "vertica-python", "duckdb"
+    "oracledb", "pyodbc", "pymssql", "trino", "clickhouse-driver", "vertica-python", "duckdb"
 ]
 
 [tool.poetry.group.dev.dependencies]