[DH-5457] Add MS SQL Server support (#422)

dataherald/db_scanner/repository/base.py

@@ -49,6 +49,7 @@ def get_all_tables_by_db(self, query: dict) -> List[TableDescription]:
     def save_table_info(self, table_info: TableDescription) -> TableDescription:
         table_info_dict = table_info.dict(exclude={"id"})
         table_info_dict["db_connection_id"] = str(table_info.db_connection_id)
+        table_info_dict["table_name"] = table_info.table_name.lower()
         table_info_dict = {
             k: v for k, v in table_info_dict.items() if v is not None and v != []
         }
@@ -57,7 +58,7 @@ def save_table_info(self, table_info: TableDescription) -> TableDescription:
                 DB_COLLECTION,
                 {
                     "db_connection_id": table_info_dict["db_connection_id"],
-                    "table_name": table_info.table_name,
+                    "table_name": table_info_dict["table_name"],
                 },
                 table_info_dict,
             )

dataherald/db_scanner/services/sql_server_scanner.py

@@ -0,0 +1,38 @@
+from overrides import override
+from sqlalchemy.exc import SQLAlchemyError
+from sqlalchemy.sql.schema import Column
+
+from dataherald.db_scanner.models.types import QueryHistory
+from dataherald.db_scanner.services.abstract_scanner import AbstractScanner
+from dataherald.sql_database.base import SQLDatabase
+
+MIN_CATEGORY_VALUE = 1
+MAX_CATEGORY_VALUE = 100
+MAX_LOGS = 5_000
+
+
+class SqlServerScanner(AbstractScanner):
+    @override
+    def cardinality_values(self, column: Column, db_engine: SQLDatabase) -> list | None:
+        try:
+            count_query = f"SELECT APPROX_COUNT_DISTINCT({column.name}) FROM {column.table.name}"  # noqa: S608
+            rs = db_engine.engine.execute(count_query).fetchall()
+        except SQLAlchemyError:
+            return None
+
+        if (
+            len(rs) > 0
+            and len(rs[0]) > 0
+            and MIN_CATEGORY_VALUE < rs[0][0] <= MAX_CATEGORY_VALUE
+        ):
+            cardinality_query = f"SELECT TOP 101 {column.name} FROM (SELECT DISTINCT {column.name} FROM [{column.table.name}]) AS subquery;"  # noqa: E501 S608
+            cardinality = db_engine.engine.execute(cardinality_query).fetchall()
+            return [str(category[0]) for category in cardinality]
+
+        return None
+
+    @override
+    def get_logs(
+        self, table: str, db_engine: SQLDatabase, db_connection_id: str  # noqa: ARG002
+    ) -> list[QueryHistory]:
+        return []

dataherald/db_scanner/sqlalchemy.py

@@ -22,6 +22,7 @@
 from dataherald.db_scanner.services.big_query_scanner import BigQueryScanner
 from dataherald.db_scanner.services.postgre_sql_scanner import PostgreSqlScanner
 from dataherald.db_scanner.services.snowflake_scanner import SnowflakeScanner
+from dataherald.db_scanner.services.sql_server_scanner import SqlServerScanner
 from dataherald.sql_database.base import SQLDatabase
 from dataherald.types import ScannerRequest
 
@@ -273,6 +274,7 @@ def scan(
             "snowflake": SnowflakeScanner,
             "bigquery": BigQueryScanner,
             "psycopg2": PostgreSqlScanner,
+            "pymssql": SqlServerScanner,
         }
         self.scanner_service = BaseScanner()
         if db_engine.engine.driver in services.keys():

dataherald/sql_database/base.py

@@ -224,7 +224,7 @@ def get_tables_and_views(self) -> List[str]:
         rows = inspector.get_table_names() + inspector.get_view_names()
         if len(rows) == 0:
             raise EmptyDBError("The db is empty it could be a permission issue")
-        return rows
+        return [row.lower() for row in rows]
 
     @property
     def dialect(self) -> str:

docs/api.create_database_connection.rst

@@ -216,6 +216,21 @@ Example::
 
 "connection_uri": mysql+pymysql://admin:[email protected]:3306/my-database
 
+Microsoft SQL Server
+^^^^^^^^^^^^
+
+Uri structure::
+
+"connection_uri": mssql+pymssql://<user>:<password>@<host>:<port>/<db-name>
+
+Example::
+
+"connection_uri": mssql+pymssql://admin:[email protected]:1433/my-database
+
+To specify a schema other than the default dbo, execute the following command::
+
+ALTER USER <your_username> WITH DEFAULT_SCHEMA = <your_schema_name>;
+
 Databricks
 ^^^^^^^^^^^^
 

requirements.txt

@@ -42,3 +42,4 @@ duckdb==0.9.1
 PyMySQL==1.1.0
 clickhouse-sqlalchemy==0.2.5
 astrapy==0.7.6
+pymssql==2.2.11