Merge pull request #22 from Snowflake-Labs/dcm-migrate-review-fixes

dcm-migrate: address PR #20 review feedback and add qualifier fixes

Author: sfc-gh-jsommerfeld

migration-tools/Cortex_Code_Skill/dcm-migrate/SKILL.md

@@ -168,14 +168,19 @@ Collect from the user:
 - Ask for the project identifier (`DB.SCHEMA.PROJECT_NAME`)
 - The project's parent DB and schema CANNOT be defined inside the project itself
 - Ask if multi-environment templating is needed
+- **Fetch the current account identifier** from the active session before writing the manifest. Run:
+  ```sql
+  SELECT CURRENT_ORGANIZATION_NAME() || '-' || CURRENT_ACCOUNT_NAME() AS ACCOUNT_IDENTIFIER
+  ```
+  Use the returned value as `account_identifier` for the target matching the current connection (typically `DEV`). For additional targets that point to other accounts (e.g. `PROD`), leave a placeholder like `<PROD_ORG>-<PROD_ACCOUNT>` and tell the user to fill it in.
 - Create local directory structure:
   ```
   <project_dir>/
   ├── manifest.yml
   └── sources/
       └── definitions/
   ```
-- Create `manifest.yml` using this template:
+- Create `manifest.yml` using this template (substitute `<ACCOUNT_IDENTIFIER>` with the value fetched above):
 
   **Minimal manifest (no templating):**
   ```yaml
@@ -185,6 +190,7 @@ Collect from the user:
 
   targets:
     DEV:
+      account_identifier: '<ACCOUNT_IDENTIFIER>'   # from CURRENT_ORGANIZATION_NAME() || '-' || CURRENT_ACCOUNT_NAME()
       project_name: 'DB_NAME.SCHEMA_NAME.PROJECT_NAME'
       project_owner: DCM_DEVELOPER
   ```
@@ -197,10 +203,12 @@ Collect from the user:
 
   targets:
     DEV:
+      account_identifier: '<ACCOUNT_IDENTIFIER>'   # current session
       project_name: 'DB_NAME.SCHEMA_NAME.PROJECT_NAME_DEV'
       project_owner: DCM_DEVELOPER
       templating_config: 'DEV'
     PROD:
+      account_identifier: '<PROD_ORG>-<PROD_ACCOUNT>'   # replace with the PROD account identifier
       project_name: 'DB_NAME.SCHEMA_NAME.PROJECT_NAME'
       project_owner: DCM_PROD_DEPLOYER
       templating_config: 'PROD'

migration-tools/Cortex_Code_Skill/dcm-migrate/scripts/ddl_to_dcm.py

@@ -107,6 +107,18 @@ def escape_jinja_conflicts(ddl_text):
 
 
 def fqn_expand(text, source_schema, object_map):
+    # Pass 1: expand SCHEMA.OBJECT references (any schema in this database)
+    # to the fully qualified "DB"."SCHEMA"."OBJECT" form. Skip if already
+    # preceded by another qualifier (e.g. "DB"."SCHEMA".NAME).
+    for target_obj in object_map:
+        t_schema = target_obj["schema"]
+        t_name = target_obj["name"]
+        t_fqn = target_obj["fqn"]
+        pattern = r'(?i)(?<!\.|")\b{}\.{}\b'.format(
+            re.escape(t_schema), re.escape(t_name)
+        )
+        text = re.sub(pattern, t_fqn, text)
+    # Pass 2: expand bare OBJECT references in the source schema.
     for target_obj in object_map:
         if target_obj["schema"] != source_schema:
             continue
@@ -538,7 +550,7 @@ def collect_future_grants(show_cmd, schema):
                 sig_for_ddl = f"{fqn}()"
 
         try:
-            res = session.sql(f"SELECT GET_DDL('{domain}', '{sig_for_ddl}') as DDL").collect()
+            res = session.sql(f"SELECT GET_DDL('{domain}', '{sig_for_ddl}', TRUE) as DDL").collect()
             ddl_text = res[0]["DDL"]
         except Exception as e:
             results.append({"schema": schema, "object_type": domain, "object_name": short_name, "status": "ERROR", "file_path": str(e)})

migration-tools/DDL_to_DCM_sproc.sql

@@ -389,6 +389,18 @@ def main(session, db_name, schema_allow_list, output_path, group_by_type):
         return full_stage_path
 
     def fqn_expand(text, source_schema):
+        # Pass 1: expand SCHEMA.OBJECT references (any schema in this database)
+        # to the fully qualified "DB"."SCHEMA"."OBJECT" form. Skip if already
+        # preceded by another qualifier (e.g. "DB"."SCHEMA".NAME).
+        for target_obj in object_map:
+            t_schema = target_obj['schema']
+            t_name = target_obj['name']
+            t_fqn = target_obj['fqn']
+            pattern = r'(?i)(?<!\.|")\b{}\.{}\b'.format(
+                re.escape(t_schema), re.escape(t_name)
+            )
+            text = re.sub(pattern, t_fqn, text)
+        # Pass 2: expand bare OBJECT references in the source schema.
         for target_obj in object_map:
             if target_obj['schema'] != source_schema:
                 continue
@@ -523,7 +535,7 @@ def main(session, db_name, schema_allow_list, output_path, group_by_type):
                 sig_for_ddl = f"{fqn}()"
 
         try:
-            res = session.sql(f"SELECT GET_DDL('{domain}', '{sig_for_ddl}') as DDL").collect()
+            res = session.sql(f"SELECT GET_DDL('{domain}', '{sig_for_ddl}', TRUE) as DDL").collect()
             ddl_text = res[0]['DDL']
         except Exception as e:
             generated_files.append((schema, domain, short_name, "ERROR", str(e)))