Structure database queries and schemas for optimal performance by avoiding nullable fields when possible, organizing query conditions efficiently, and using targeted indexes. **Key practices:** 1. **Avoid nullable fields for performance**: Use default values instead of nullable fields, especially in ClickHouse where non-null values perform better: ```typescript // Preferred severity: { type: CHLowCardinality(CHString('none')) } critical: { type: CHBoolean(false) } // Instead of severity: { type: CHNullable(CHString()) } critical: { type: CHNullable(CHBoolean()) } ``` 2. **Structure conditions to minimize OR operations**: Group related conditions into arrays rather than chaining multiple OR statements: ```typescript // Preferred - cleaner and more performant const snoozedCondition: Array = []; if (query.snoozed === false) { snoozedCondition.push({ snoozedUntil: { $exists: false } }, { snoozedUntil: null }); } // Instead of multiple OR statements in the main query requestQuery.$or = [{ snoozedUntil: { $exists: false } }, { snoozedUntil: null }]; ``` 3. **Use partial indexes for filtered queries**: Apply partial filter expressions to create more efficient indexes for specific use cases: ```typescript // Targeted index with partial filtering messageSchema.index( { _subscriberId: 1, _environmentId: 1, read: 1, archived: 1, seen: 1, createdAt: -1, }, { name: 'in_app_messages_count', partialFilterExpression: { channel: 'in_app' }, } ); ``` This approach improves query performance, reduces index size, and ensures better database resource utilization across different database systems.