feat: optimize _index_query performance in OceanBase Oracle mode

[yuanoOo]

🚀 Feature Description

Optimize the performance of _index_query method in OceanBase Oracle mode by adding early filtering conditions to reduce data scanning.

🎯 Optimization Details

• Override _index_query method in OceanBaseDialect_cx_oracle class
• Add filtering condition on all_ind_columns.table_name
• Filter data early in query execution to avoid full table scan

🔧 Technical Implementation

Key Optimization

# Added in WHERE clause:
dictionary.all_ind_columns.c.table_name.in_(bindparam("all_objects"))

SQL Query Optimization

Before:

WHERE a_indexes.table_owner = :owner 
  AND a_indexes.table_name IN (:all_objects)

After:

WHERE a_indexes.table_owner = :owner 
  AND a_indexes.table_name IN (:all_objects)
  AND a_ind_columns.table_name IN (:all_objects)  -- New optimization condition

📈 Performance Benefits

• ✅ Reduces data scanning on all_ind_columns table
• ✅ Filters target table data early in query execution
• ✅ Avoids full table scan, significantly improving query performance
• ✅ Particularly beneficial for databases with large numbers of indexes

🧪 Testing & Validation

• ✅ Created and ran test script to verify optimization is effective
• ✅ Confirmed generated SQL contains optimization conditions
• ✅ Compatible with SQLAlchemy 2.x version

📝 Compatibility

• Only supports SQLAlchemy 2.x version
• Version control through SA_20_PLUS flag
• No impact on existing functionality

🔗 Related Files

• oceanbase-sqlalchemy-plugin/oceanbase_sqlalchemy/cx_oracle.py