Data Preview

Why these restrictions:

• Prevents data modification
• Protects system integrity
• Enforces SAP security model
• Safe for AI-generated queries
• Read-only exploration

Supported Object Types

Required Parameters

• object_name: Name of table or CDS view to query
• sql_query: SELECT statement (validated for safety)

Optional Parameters

• object_type: ddic (default) or cds
• max_rows: Maximum rows to return (1-1000, default: 100)

SQL Query Validation

Automatic validation checks:

1. ✅ Query starts with SELECT
2. ✅ No DML keywords (INSERT, UPDATE, DELETE)
3. ✅ No DDL keywords (CREATE, DROP, ALTER)
4. ✅ No SQL comments (--, /*, */)
5. ✅ Single statement only (no semicolons except at end)
6. ✅ Query references the specified object

Rejected query examples:

❌ "SELECT * FROM mara; DELETE FROM mara"
❌ "INSERT INTO table VALUES (...)"
❌ "UPDATE table SET field = value"
❌ "DROP TABLE ztable"
❌ "SELECT * FROM mara -- comment"

Accepted query examples:

✅ "SELECT * FROM mara"
✅ "SELECT matnr, mtart FROM mara WHERE mtart = 'FERT'"
✅ "SELECT TOP 10 * FROM vbak"
✅ "SELECT * FROM I_SalesOrder WHERE SalesOrderType = 'OR'"

Output Format

Returns structured JSON:

{
  "success": true,
  "object_name": "MARA",
  "object_type": "ddic",
  "row_count": 15,
  "max_rows": 100,
  "columns": [
    {
      "name": "MATNR",
      "type": "CHAR",
      "length": 40,
      "description": "Material Number"
    }
  ],
  "rows": [
    {
      "MATNR": "000000000000000001",
      "MTART": "FERT",
      "MATKL": "001"
    }
  ]
}

Usage Examples

Preview table contents:

Show me the first 10 rows from table MARA

Query with filter:

Get all sales orders from table VBAK where order type is 'OR'

CDS view query:

Query CDS view I_SalesOrder for orders created in 2024

Specific fields:

Show material number and description from MARA for finished goods

Join exploration:

Query ZI_BOOKING to see customer and flight associations

Data validation:

After creating ZI_CUSTOMER, preview the data to verify associations work

Common Query Patterns

Basic Preview

SELECT * FROM table_name
Purpose: Quick overview of table structure and data

Limited Rows

SELECT TOP 10 * FROM table_name
Purpose: Sample data without overwhelming output

Specific Fields

SELECT field1, field2, field3 FROM table_name
Purpose: Focus on relevant columns

Filtered Data

SELECT * FROM table_name WHERE field = 'value'
Purpose: Narrow down to relevant records

Ordered Results

SELECT * FROM table_name ORDER BY field DESC
Purpose: See newest/highest/sorted data first

CDS View with Parameters

SELECT * FROM cds_view( parameter = 'value' )
Purpose: Test CDS view parameters

Workflow Integration

Recommended Workflow

This is step 4 in the research-query workflow:

1. Use SAP documentation tools → Identify relevant table/CDS name

   Search SAP Help for "sales order data model"
   Result: VBAK, VBAP tables or I_SalesOrder CDS
   

2. Use GetObjectInfo → Retrieve metadata/structure

   Get object info for I_SalesOrder
   Result: Field definitions, associations, annotations
   

3. AI formulates SQL query → Based on structure and requirement

   AI: "Based on fields, query should be:
        SELECT SalesOrder, Customer, CreationDate 
        FROM I_SalesOrder 
        WHERE CreationDate >= '2024-01-01'"
   

4. Execute with data_preview → Run the query

   Execute query on I_SalesOrder
   Result: Actual data matching criteria
   

Development Workflow

1. create_ai_object → Create CDS view
2. activate_object → Activate and check syntax
3. data_preview → Test CDS view returns expected data
4. Iterate if needed

Exploration Workflow

1. search_object → Find relevant tables/views
2. get_object_info → Understand structure
3. data_preview → Explore actual data
4. Use insights for development

Tips for Effective Queries

1. Start with LIMIT: Use TOP 10 for initial exploration
2. Specific fields: Don't always SELECT * - choose relevant fields
3. Add filters: WHERE clauses reduce result size
4. Test incrementally: Start simple, add complexity
5. Mind max_rows: Default 100, increase if needed
6. Use field names carefully: SAP field names are case-sensitive in some contexts
7. Test CDS parameters: Verify parameter handling works
8. Check associations: Query associated data to test joins

Output Expectations

When Copilot uses this skill, expect:

• Row count and data sample
• Column definitions with types
• Formatted data table
• Indication if results truncated (due to max_rows)
• Data interpretation if relevant
• Next steps: "Modify query to filter further" or "Data looks correct"

Integration with Other Skills

Data exploration workflow:

1. search_object → Find tables/views
2. get_object_info → Understand structure
3. data_preview → See actual data
4. Use findings in development

CDS development workflow:

1. sap_help_search → Research CDS patterns
2. create_ai_object → Create CDS view
3. data_preview → Test CDS view
4. change_ai_object → Fix issues if found
5. Re-run data_preview to verify

Validation workflow:

1. create_ai_object → Create object
2. get_atc_results → Check code quality
3. data_preview → Validate data correctness
4. change_ai_object if issues found

Common Use Cases

1. Table Content Preview

Query: "SELECT TOP 100 * FROM mara"
Use: See what material master data looks like

2. CDS View Testing

Query: "SELECT * FROM ZI_BOOKING"
Use: Verify newly created CDS view returns data

3. Filter Validation

Query: "SELECT * FROM vbak WHERE erdat >= '20240101'"
Use: Test date filter works correctly

4. Association Check

Query: "SELECT booking_id, customer_id, customer._name FROM ZI_BOOKING"
Use: Verify CDS association to customer works

5. Data Pattern Analysis

Query: "SELECT DISTINCT mtart FROM mara"
Use: Find all material types in system

6. Record Count

Query: "SELECT COUNT(*) as record_count FROM table"
Use: Check table size

Error Handling

Common errors and fixes:

Security & Authorization

SAP authorization applies:

• You can only query tables/views you have authority for
• Authorization object S_TABU_DIS controls table access
• CDS view access controls (DCL) are enforced
• User's existing authorities determine what data is visible

Authorization errors:

Error: No authorization for table MARA
Solution: Request authorization from SAP Basis team
Workaround: Use CDS views with appropriate DCL

SAP SQL Dialect

SAP HANA SQL features (S/4HANA):

• TOP clause for limiting rows
• Advanced functions (STRING_AGG, etc.)
• CDS view parameters
• Associations in SELECT

Traditional SQL features (ECC):

• Standard WHERE, ORDER BY
• Aggregate functions (COUNT, SUM, AVG)
• JOIN operations
• DISTINCT

Consult SAP documentation for specific SQL capabilities of your system.

Best Practices

1. Preview before building: Check data before creating CDS views
2. Use TOP/LIMIT: Don't query millions of rows
3. Specific fields: SELECT specific fields, not always *
4. Test associations: When creating CDS, test associations with data_preview
5. Iterate queries: Start simple, refine based on results
6. Check field names: Use get_object_info to verify correct field names
7. Mind performance: Even SELECT can impact system if not careful
8. Respect authorizations: Don't try to bypass SAP security

Limitations

What this tool CANNOT do:

• ❌ Modify data (by design - security feature)
• ❌ Create/drop objects (use create_ai_object instead)
• ❌ Execute stored procedures (SELECT only)
• ❌ Access tables user has no authority for
• ❌ Bypass SAP security model

When you need data modification:

• Use SAP GUI transactions
• Create ABAP programs with proper authority
• Use SAP standard BAPIs/APIs
• This tool is intentionally read-only for safety

Data Privacy

Important reminders:

• Respect data privacy laws (GDPR, etc.)
• Don't query sensitive personal data unnecessarily
• Be aware of what data you're viewing
• Follow your organization's data policies
• Use WHERE clauses to limit personal data exposure
• Consider masking/anonymization requirements