From 71492f1fb83727868e64cdaad0c5b5caea76b5ab Mon Sep 17 00:00:00 2001
From: Sean Preston <sean@ethyca.com>
Date: Fri, 3 Jun 2022 08:20:35 -0400
Subject: [PATCH] [#557] MSSQL discovery script (#581)

* adds script to discover mssql datastore compatibility

* make prints consistent

* updates changelog

* add URL template

* add warning

* store columns correctly

* move uncomitted secrets to another file to add to .gitignore

* remove empty secrets file, add to gitignore

* add comment explaining lack of secrets
---
 .gitignore                |  5 ++-
 CHANGELOG.md              |  3 ++
 scripts/mssql_discover.py | 79 +++++++++++++++++++++++++++++++++++++++
 3 files changed, 86 insertions(+), 1 deletion(-)
 create mode 100644 scripts/mssql_discover.py

diff --git a/.gitignore b/.gitignore
index 19ec29a7d..c2d8e5c1e 100644
--- a/.gitignore
+++ b/.gitignore
@@ -138,4 +138,7 @@ envfiles/
 fides_uploads
 
 # Prevent SaaS configs from being committed
-saas_config.toml
\ No newline at end of file
+saas_config.toml
+
+# Script secrets
+scripts/secrets.py
\ No newline at end of file
diff --git a/CHANGELOG.md b/CHANGELOG.md
index 0e9ae4ea4..42398565d 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -31,6 +31,9 @@ The types of changes are:
 * Use the `RuleResponse` schema within the `PrivacyRequestReposnse` schema [#580](https://github.com/ethyca/fidesops/pull/580)
 * Updated the webserver to use `PORT` config variable from the `fidesops.toml` file [#586](https://github.com/ethyca/fidesops/pull/586)
 
+### Developer Experience
+* Adds a script for MSSQL schema exploration [#557](https://github.com/ethyca/fidesops/pull/581)
+
 ## [1.5.1](https://github.com/ethyca/fidesops/compare/1.5.0...1.5.1) - 2022-05-27
 
 ### Added
diff --git a/scripts/mssql_discover.py b/scripts/mssql_discover.py
new file mode 100644
index 000000000..1ac047d61
--- /dev/null
+++ b/scripts/mssql_discover.py
@@ -0,0 +1,79 @@
+import sqlalchemy
+
+# This file is not committed to the repo, please create secrets.py with the required
+# variables in the same dir as this file before running this script
+from secrets import (
+    USER,
+    PASS,
+    IP,
+    PORT,
+    DB,
+)
+
+MASTER_MSSQL_URL = f"mssql+pyodbc://{USER}:{PASS}@{IP}:{PORT}/{DB}?driver=ODBC+Driver+17+for+SQL+Server"
+
+
+SUPPORTED_DATA_TYPES = set(
+    [
+        # char types
+        "varchar",
+        "nvarchar",
+        "char",
+        "nchar",
+        "ntext",
+        "text",
+        # numeric types
+        "int",
+        "bigint",
+        "smallint",
+        "tinyint",
+        "money",
+        "float",
+        "decimal",
+        # date types
+        "date",
+        "datetime",
+        "datetime2",
+        "smalldatetime",
+        # other types
+        "bit",
+    ]
+)
+
+
+def mssql_discover():
+    """
+    Select all databases from the instance
+    Select the schema data for each data base
+    Check if there are any fields in the schema that Fidesops does not yet support
+    """
+    engine = sqlalchemy.create_engine(MASTER_MSSQL_URL)
+    all_dbs = engine.execute("SELECT name FROM sys.databases;").all()
+    all_columns = []
+    flagged_columns = []
+    flagged_datatypes = set()
+    for db_name in all_dbs:
+        db_name = db_name[0]
+        try:
+            columns = engine.execute(
+                f"SELECT TABLE_NAME, COLUMN_NAME, DATA_TYPE FROM {db_name}.INFORMATION_SCHEMA.COLUMNS;"
+            ).all()
+        except Exception:
+            continue
+
+        all_columns.extend(columns)
+        for table, column, data_type in columns:
+            if data_type not in SUPPORTED_DATA_TYPES:
+                flagged_datatypes.add(data_type)
+                flagged_columns.append(f"{db_name}.{table}.{column}: {data_type}")
+
+    print(f"{len(all_columns)} columns found")
+    print(f"{len(flagged_columns)} columns flagged")
+    print(f"Flagged datatypes:")
+    print(",\n".join(flagged_datatypes))
+    print(f"Flagged columns:")
+    print(",\n".join(flagged_columns))
+
+
+if __name__ == "__main__":
+    mssql_discover()