rez_core 7.1.21 → 7.1.22

package/docs/EAV_FILTER_IMPLEMENTATION.md

@@ -1,420 +1,420 @@
-# EAV Filter Implementation Guide
-
-## Overview
-
-The filter service now supports filtering on both **DB_COL** (database columns) and **EAV** (Entity-Attribute-Value) attributes. This allows you to filter entities based on attributes stored in either the main entity table or in separate EAV tables.
-
-## Architecture
-
-### Key Components
-
-1. **getEavTableName()** - Maps element_type to the appropriate EAV table
-2. **buildWhereClauses()** - Separates filters into DB_COL and EAV categories
-3. **buildEavCondition()** - Builds WHERE conditions for EAV attributes
-4. **EAV Helper Methods** - Type-specific condition builders for EAV attributes
-
-### Data Flow
-
-```
-Filter Request
-    ↓
-buildWhereClauses()
-    ↓
-Split into:
-    ├─→ DB_COL Clauses (for regular columns)
-    ├─→ EAV Joins (for EAV table LEFT JOINs)
-    └─→ EAV Clauses (for EAV WHERE conditions)
-    ↓
-Query Builder
-    ├─→ Add EAV LEFT JOINs
-    ├─→ Apply DB_COL WHERE clauses
-    └─→ Apply EAV WHERE clauses
-    ↓
-Execute Query
-```
-
-## EAV Table Mapping
-
-The system automatically maps element types to EAV tables:
-
-| Element Type | EAV Table |
-|-------------|-----------|
-| text, label, textarea, email, phone, password, url | frm_eav_varchar |
-| number, file-upload | frm_eav_int |
-| decimal | frm_eav_decimal |
-| date | frm_eav_date |
-| time | frm_eav_time |
-| datetime | frm_eav_timestamp |
-| checkbox | frm_eav_boolean |
-| text-editor, multiselect | frm_eav_text |
-| json-editor | frm_eav_json |
-
-## How It Works
-
-### 1. Attribute Detection
-
-When a filter is applied, the system checks the `storage_type` field in the `AttributeMaster` table:
-
-```typescript
-const isEav = meta.storage_type?.toUpperCase() === 'EAV';
-```
-
-### 2. EAV Join Creation
-
-For each EAV attribute filter, a LEFT JOIN is created:
-
-```sql
-LEFT JOIN frm_eav_varchar AS eav_father_sal_0
-  ON eav_father_sal_0.entity_type = 'LFMG'
-  AND eav_father_sal_0.entity_id = CAST(e.id AS VARCHAR)
-  AND eav_father_sal_0.key = 'father_sal'
-```
-
-### 3. EAV Condition Application
-
-WHERE conditions are applied on the joined EAV table:
-
-```sql
-WHERE eav_father_sal_0.value = '100'
-```
-
-## Usage Examples
-
-### Example 1: Filter Lead by EAV Attribute
-
-```typescript
-// Filter Lead-fmg by EAV attribute "father_sal"
-const filterRequest: FilterRequestDto = {
-  entity_type: 'LEAD',
-  quickFilter: [
-    {
-      filter_attribute: 'father_sal',
-      filter_operator: 'equal',
-      filter_value: 100,
-      filter_entity_type: 'LFMG'
-    }
-  ],
-  loggedInUser: userData
-};
-
-const result = await filterService.applyFilterWrapper(filterRequest);
-```
-
-### Example 2: Mixed DB_COL and EAV Filters
-
-```typescript
-// Filter with both DB_COL and EAV attributes
-const filterRequest: FilterRequestDto = {
-  entity_type: 'LEAD',
-  quickFilter: [
-    // DB_COL filter
-    {
-      filter_attribute: 'name',
-      filter_operator: 'contains',
-      filter_value: 'John',
-      filter_entity_type: 'LEAD'
-    },
-    // EAV filter on sub-entity
-    {
-      filter_attribute: 'father_sal',
-      filter_operator: 'greater_than',
-      filter_value: 50,
-      filter_entity_type: 'LFMG'
-    },
-    // EAV filter on sub-entity
-    {
-      filter_attribute: 'father_nick_name',
-      filter_operator: 'contains',
-      filter_value: 'Dad',
-      filter_entity_type: 'LFMG'
-    }
-  ],
-  loggedInUser: userData
-};
-```
-
-### Example 3: Filter by Date EAV Attribute
-
-```typescript
-const filterRequest: FilterRequestDto = {
-  entity_type: 'LEAD',
-  quickFilter: [
-    {
-      filter_attribute: 'birth_date',
-      filter_operator: 'after',
-      filter_value: '2000-01-01',
-      filter_entity_type: 'LFMG'
-    }
-  ],
-  loggedInUser: userData
-};
-```
-
-## Supported Operators by EAV Type
-
-### Text EAV (varchar, text)
-- `contains` - Partial match (case-insensitive)
-- `equal` - Exact match
-- `not_equal` - Not equal
-- `empty` - NULL check
-- `not_empty` - NOT NULL check
-
-### Number EAV (int, decimal)
-- `equal`
-- `not_equal`
-- `greater_than`
-- `less_than`
-- `less_than_qual_to`
-- `greater_than_qual_to`
-- `empty`
-- `not_empty`
-
-### Date EAV (date, timestamp)
-- `equal` / `is`
-- `before` / `is_before`
-- `after` / `is_after`
-- `is_on_or_before`
-- `is_on_or_after`
-- `between`
-- `empty`
-- `not_empty`
-
-### Boolean EAV (boolean)
-- `equal`
-- `not_equal`
-- `empty`
-- `not_empty`
-
-### Multiselect EAV (text, json)
-- `equal` / `contains`
-- `not_equal` / `not_contains`
-- `empty`
-- `not_empty`
-
-## Performance Considerations
-
-### LEFT JOIN Approach
-
-The implementation uses LEFT JOIN for EAV attributes, which:
-- ✅ Allows filtering by EAV attributes
-- ✅ Returns all main entity records that match
-- ✅ Handles NULL values correctly
-- ⚠️ May impact performance with many EAV filters (multiple JOINs)
-
-### Optimization Tips
-
-1. **Minimize EAV Filters**: Use DB_COL attributes when possible
-2. **Index EAV Tables**: Ensure indexes on `(entity_type, entity_id, key)` in EAV tables
-3. **Limit Filter Count**: Too many EAV filters = too many JOINs
-4. **Use Specific Operators**: Avoid `contains` on large datasets
-
-## Sub-Entity Filtering
-
-The filter service supports filtering the main entity based on sub-entity EAV attributes through the `applyFilterWrapper` method:
-
-### How It Works
-
-1. Filters are grouped by `filter_entity_type`
-2. Sub-entity filters are applied first to get matching sub-entity IDs
-3. Related main entity IDs are retrieved via `entity-relation-data`
-4. Main entity is filtered using the intersection of all related IDs
-5. Main entity filters (DB_COL and EAV) are then applied
-
-### Example: Filter Lead by Lead-fmg EAV Attributes
-
-```typescript
-const filterRequest: FilterRequestDto = {
-  entity_type: 'LEAD', // Main entity
-  quickFilter: [
-    {
-      filter_attribute: 'father_sal',
-      filter_operator: 'equal',
-      filter_value: 100,
-      filter_entity_type: 'LFMG' // Sub-entity
-    },
-    {
-      filter_attribute: 'father_nick_name',
-      filter_operator: 'contains',
-      filter_value: 'John',
-      filter_entity_type: 'LFMG' // Sub-entity
-    }
-  ],
-  loggedInUser: userData
-};
-
-// This will:
-// 1. Find all LFMG records where father_sal = 100 AND father_nick_name LIKE '%John%'
-// 2. Get related LEAD IDs from entity-relation-data
-// 3. Return LEAD records with those IDs
-const result = await filterService.applyFilterWrapper(filterRequest);
-```
-
-## Query Examples
-
-### Generated SQL for EAV Filter
-
-**Input:**
-```json
-{
-  "entity_type": "LEAD",
-  "quickFilter": [
-    {
-      "filter_attribute": "father_sal",
-      "filter_operator": "greater_than",
-      "filter_value": 50,
-      "filter_entity_type": "LFMG"
-    }
-  ]
-}
-```
-
-**Generated SQL:**
-```sql
-SELECT e.*
-FROM rez.frm_lead AS e
-LEFT JOIN rez.frm_eav_int AS eav_father_sal_0
-  ON eav_father_sal_0.entity_type = 'LFMG'
-  AND eav_father_sal_0.entity_id = CAST(e.id AS VARCHAR)
-  AND eav_father_sal_0.key = 'father_sal'
-WHERE eav_father_sal_0.value > '50'
-  AND e.level_type = 'ORG'
-  AND e.level_id = '123'
-```
-
-## Troubleshooting
-
-### Issue: EAV filter returns no results
-
-**Possible Causes:**
-1. No EAV records exist for the entity
-2. Incorrect `entity_type` in EAV table
-3. Wrong `key` value in EAV table
-4. Data type mismatch
-
-**Solution:**
-```sql
--- Check if EAV records exist
-SELECT * FROM frm_eav_varchar
-WHERE entity_type = 'LFMG'
-  AND entity_id = '123'
-  AND key = 'father_sal';
-```
-
-### Issue: Performance degradation with many filters
-
-**Solution:**
-- Add indexes on EAV tables:
-```sql
-CREATE INDEX idx_eav_varchar_lookup
-ON frm_eav_varchar (entity_type, entity_id, key);
-```
-
-### Issue: Sub-entity filter returns empty
-
-**Solution:**
-- Verify entity relations exist in `entity-relation-data`
-```sql
-SELECT * FROM entity_relation_data
-WHERE source_entity_type = 'LEAD'
-  AND target_entity_type = 'LFMG';
-```
-
-## API Changes
-
-### Updated Methods
-
-1. **buildWhereClauses()** - Now returns object with three properties:
-   ```typescript
-   {
-     dbColClauses: Array<{query: string, params: Record<string, any>}>,
-     eavJoins: Array<{alias: string, tableName: string, entityType: string, attributeKey: string}>,
-     eavClauses: Array<{query: string, params: Record<string, any>}>
-   }
-   ```
-
-2. **applyFilter()** - Automatically handles EAV joins and clauses
-3. **GetFilterData()** - Updated to support EAV filtering
-
-### No Breaking Changes
-
-All existing API signatures remain the same. The EAV support is transparent to consumers of the FilterService.
-
-## Testing
-
-### Test Case 1: EAV Text Filter
-
-```typescript
-it('should filter by EAV text attribute', async () => {
-  const result = await filterService.applyFilter({
-    entity_type: 'LEAD',
-    quickFilter: [{
-      filter_attribute: 'father_nick_name',
-      filter_operator: 'contains',
-      filter_value: 'John',
-      filter_entity_type: 'LFMG'
-    }],
-    loggedInUser: userData
-  });
-
-  expect(result.success).toBe(true);
-  expect(result.data.entity_list.length).toBeGreaterThan(0);
-});
-```
-
-### Test Case 2: EAV Number Filter
-
-```typescript
-it('should filter by EAV number attribute', async () => {
-  const result = await filterService.applyFilter({
-    entity_type: 'LEAD',
-    quickFilter: [{
-      filter_attribute: 'father_sal',
-      filter_operator: 'greater_than',
-      filter_value: 100,
-      filter_entity_type: 'LFMG'
-    }],
-    loggedInUser: userData
-  });
-
-  expect(result.success).toBe(true);
-});
-```
-
-### Test Case 3: Mixed DB_COL and EAV
-
-```typescript
-it('should filter by both DB_COL and EAV attributes', async () => {
-  const result = await filterService.applyFilter({
-    entity_type: 'LEAD',
-    quickFilter: [
-      {
-        filter_attribute: 'name', // DB_COL
-        filter_operator: 'contains',
-        filter_value: 'Lead',
-        filter_entity_type: 'LEAD'
-      },
-      {
-        filter_attribute: 'father_sal', // EAV
-        filter_operator: 'equal',
-        filter_value: 100,
-        filter_entity_type: 'LFMG'
-      }
-    ],
-    loggedInUser: userData
-  });
-
-  expect(result.success).toBe(true);
-});
-```
-
-## Summary
-
-The EAV filter implementation provides:
-- ✅ Seamless filtering on both DB_COL and EAV attributes
-- ✅ Support for all standard operators (equal, contains, greater_than, etc.)
-- ✅ Sub-entity filtering with automatic relation resolution
-- ✅ Type-safe condition building for different data types
-- ✅ No breaking changes to existing API
-- ✅ Automatic EAV table detection and JOIN generation
-
-The system is production-ready and can handle complex filtering scenarios involving main entities and multiple sub-entities with mixed storage types (DB_COL and EAV).
+# EAV Filter Implementation Guide
+
+## Overview
+
+The filter service now supports filtering on both **DB_COL** (database columns) and **EAV** (Entity-Attribute-Value) attributes. This allows you to filter entities based on attributes stored in either the main entity table or in separate EAV tables.
+
+## Architecture
+
+### Key Components
+
+1. **getEavTableName()** - Maps element_type to the appropriate EAV table
+2. **buildWhereClauses()** - Separates filters into DB_COL and EAV categories
+3. **buildEavCondition()** - Builds WHERE conditions for EAV attributes
+4. **EAV Helper Methods** - Type-specific condition builders for EAV attributes
+
+### Data Flow
+
+```
+Filter Request
+    ↓
+buildWhereClauses()
+    ↓
+Split into:
+    ├─→ DB_COL Clauses (for regular columns)
+    ├─→ EAV Joins (for EAV table LEFT JOINs)
+    └─→ EAV Clauses (for EAV WHERE conditions)
+    ↓
+Query Builder
+    ├─→ Add EAV LEFT JOINs
+    ├─→ Apply DB_COL WHERE clauses
+    └─→ Apply EAV WHERE clauses
+    ↓
+Execute Query
+```
+
+## EAV Table Mapping
+
+The system automatically maps element types to EAV tables:
+
+| Element Type | EAV Table |
+|-------------|-----------|
+| text, label, textarea, email, phone, password, url | frm_eav_varchar |
+| number, file-upload | frm_eav_int |
+| decimal | frm_eav_decimal |
+| date | frm_eav_date |
+| time | frm_eav_time |
+| datetime | frm_eav_timestamp |
+| checkbox | frm_eav_boolean |
+| text-editor, multiselect | frm_eav_text |
+| json-editor | frm_eav_json |
+
+## How It Works
+
+### 1. Attribute Detection
+
+When a filter is applied, the system checks the `storage_type` field in the `AttributeMaster` table:
+
+```typescript
+const isEav = meta.storage_type?.toUpperCase() === 'EAV';
+```
+
+### 2. EAV Join Creation
+
+For each EAV attribute filter, a LEFT JOIN is created:
+
+```sql
+LEFT JOIN frm_eav_varchar AS eav_father_sal_0
+  ON eav_father_sal_0.entity_type = 'LFMG'
+  AND eav_father_sal_0.entity_id = CAST(e.id AS VARCHAR)
+  AND eav_father_sal_0.key = 'father_sal'
+```
+
+### 3. EAV Condition Application
+
+WHERE conditions are applied on the joined EAV table:
+
+```sql
+WHERE eav_father_sal_0.value = '100'
+```
+
+## Usage Examples
+
+### Example 1: Filter Lead by EAV Attribute
+
+```typescript
+// Filter Lead-fmg by EAV attribute "father_sal"
+const filterRequest: FilterRequestDto = {
+  entity_type: 'LEAD',
+  quickFilter: [
+    {
+      filter_attribute: 'father_sal',
+      filter_operator: 'equal',
+      filter_value: 100,
+      filter_entity_type: 'LFMG'
+    }
+  ],
+  loggedInUser: userData
+};
+
+const result = await filterService.applyFilterWrapper(filterRequest);
+```
+
+### Example 2: Mixed DB_COL and EAV Filters
+
+```typescript
+// Filter with both DB_COL and EAV attributes
+const filterRequest: FilterRequestDto = {
+  entity_type: 'LEAD',
+  quickFilter: [
+    // DB_COL filter
+    {
+      filter_attribute: 'name',
+      filter_operator: 'contains',
+      filter_value: 'John',
+      filter_entity_type: 'LEAD'
+    },
+    // EAV filter on sub-entity
+    {
+      filter_attribute: 'father_sal',
+      filter_operator: 'greater_than',
+      filter_value: 50,
+      filter_entity_type: 'LFMG'
+    },
+    // EAV filter on sub-entity
+    {
+      filter_attribute: 'father_nick_name',
+      filter_operator: 'contains',
+      filter_value: 'Dad',
+      filter_entity_type: 'LFMG'
+    }
+  ],
+  loggedInUser: userData
+};
+```
+
+### Example 3: Filter by Date EAV Attribute
+
+```typescript
+const filterRequest: FilterRequestDto = {
+  entity_type: 'LEAD',
+  quickFilter: [
+    {
+      filter_attribute: 'birth_date',
+      filter_operator: 'after',
+      filter_value: '2000-01-01',
+      filter_entity_type: 'LFMG'
+    }
+  ],
+  loggedInUser: userData
+};
+```
+
+## Supported Operators by EAV Type
+
+### Text EAV (varchar, text)
+- `contains` - Partial match (case-insensitive)
+- `equal` - Exact match
+- `not_equal` - Not equal
+- `empty` - NULL check
+- `not_empty` - NOT NULL check
+
+### Number EAV (int, decimal)
+- `equal`
+- `not_equal`
+- `greater_than`
+- `less_than`
+- `less_than_qual_to`
+- `greater_than_qual_to`
+- `empty`
+- `not_empty`
+
+### Date EAV (date, timestamp)
+- `equal` / `is`
+- `before` / `is_before`
+- `after` / `is_after`
+- `is_on_or_before`
+- `is_on_or_after`
+- `between`
+- `empty`
+- `not_empty`
+
+### Boolean EAV (boolean)
+- `equal`
+- `not_equal`
+- `empty`
+- `not_empty`
+
+### Multiselect EAV (text, json)
+- `equal` / `contains`
+- `not_equal` / `not_contains`
+- `empty`
+- `not_empty`
+
+## Performance Considerations
+
+### LEFT JOIN Approach
+
+The implementation uses LEFT JOIN for EAV attributes, which:
+- ✅ Allows filtering by EAV attributes
+- ✅ Returns all main entity records that match
+- ✅ Handles NULL values correctly
+- ⚠️ May impact performance with many EAV filters (multiple JOINs)
+
+### Optimization Tips
+
+1. **Minimize EAV Filters**: Use DB_COL attributes when possible
+2. **Index EAV Tables**: Ensure indexes on `(entity_type, entity_id, key)` in EAV tables
+3. **Limit Filter Count**: Too many EAV filters = too many JOINs
+4. **Use Specific Operators**: Avoid `contains` on large datasets
+
+## Sub-Entity Filtering
+
+The filter service supports filtering the main entity based on sub-entity EAV attributes through the `applyFilterWrapper` method:
+
+### How It Works
+
+1. Filters are grouped by `filter_entity_type`
+2. Sub-entity filters are applied first to get matching sub-entity IDs
+3. Related main entity IDs are retrieved via `entity-relation-data`
+4. Main entity is filtered using the intersection of all related IDs
+5. Main entity filters (DB_COL and EAV) are then applied
+
+### Example: Filter Lead by Lead-fmg EAV Attributes
+
+```typescript
+const filterRequest: FilterRequestDto = {
+  entity_type: 'LEAD', // Main entity
+  quickFilter: [
+    {
+      filter_attribute: 'father_sal',
+      filter_operator: 'equal',
+      filter_value: 100,
+      filter_entity_type: 'LFMG' // Sub-entity
+    },
+    {
+      filter_attribute: 'father_nick_name',
+      filter_operator: 'contains',
+      filter_value: 'John',
+      filter_entity_type: 'LFMG' // Sub-entity
+    }
+  ],
+  loggedInUser: userData
+};
+
+// This will:
+// 1. Find all LFMG records where father_sal = 100 AND father_nick_name LIKE '%John%'
+// 2. Get related LEAD IDs from entity-relation-data
+// 3. Return LEAD records with those IDs
+const result = await filterService.applyFilterWrapper(filterRequest);
+```
+
+## Query Examples
+
+### Generated SQL for EAV Filter
+
+**Input:**
+```json
+{
+  "entity_type": "LEAD",
+  "quickFilter": [
+    {
+      "filter_attribute": "father_sal",
+      "filter_operator": "greater_than",
+      "filter_value": 50,
+      "filter_entity_type": "LFMG"
+    }
+  ]
+}
+```
+
+**Generated SQL:**
+```sql
+SELECT e.*
+FROM rez.frm_lead AS e
+LEFT JOIN rez.frm_eav_int AS eav_father_sal_0
+  ON eav_father_sal_0.entity_type = 'LFMG'
+  AND eav_father_sal_0.entity_id = CAST(e.id AS VARCHAR)
+  AND eav_father_sal_0.key = 'father_sal'
+WHERE eav_father_sal_0.value > '50'
+  AND e.level_type = 'ORG'
+  AND e.level_id = '123'
+```
+
+## Troubleshooting
+
+### Issue: EAV filter returns no results
+
+**Possible Causes:**
+1. No EAV records exist for the entity
+2. Incorrect `entity_type` in EAV table
+3. Wrong `key` value in EAV table
+4. Data type mismatch
+
+**Solution:**
+```sql
+-- Check if EAV records exist
+SELECT * FROM frm_eav_varchar
+WHERE entity_type = 'LFMG'
+  AND entity_id = '123'
+  AND key = 'father_sal';
+```
+
+### Issue: Performance degradation with many filters
+
+**Solution:**
+- Add indexes on EAV tables:
+```sql
+CREATE INDEX idx_eav_varchar_lookup
+ON frm_eav_varchar (entity_type, entity_id, key);
+```
+
+### Issue: Sub-entity filter returns empty
+
+**Solution:**
+- Verify entity relations exist in `entity-relation-data`
+```sql
+SELECT * FROM entity_relation_data
+WHERE source_entity_type = 'LEAD'
+  AND target_entity_type = 'LFMG';
+```
+
+## API Changes
+
+### Updated Methods
+
+1. **buildWhereClauses()** - Now returns object with three properties:
+   ```typescript
+   {
+     dbColClauses: Array<{query: string, params: Record<string, any>}>,
+     eavJoins: Array<{alias: string, tableName: string, entityType: string, attributeKey: string}>,
+     eavClauses: Array<{query: string, params: Record<string, any>}>
+   }
+   ```
+
+2. **applyFilter()** - Automatically handles EAV joins and clauses
+3. **GetFilterData()** - Updated to support EAV filtering
+
+### No Breaking Changes
+
+All existing API signatures remain the same. The EAV support is transparent to consumers of the FilterService.
+
+## Testing
+
+### Test Case 1: EAV Text Filter
+
+```typescript
+it('should filter by EAV text attribute', async () => {
+  const result = await filterService.applyFilter({
+    entity_type: 'LEAD',
+    quickFilter: [{
+      filter_attribute: 'father_nick_name',
+      filter_operator: 'contains',
+      filter_value: 'John',
+      filter_entity_type: 'LFMG'
+    }],
+    loggedInUser: userData
+  });
+
+  expect(result.success).toBe(true);
+  expect(result.data.entity_list.length).toBeGreaterThan(0);
+});
+```
+
+### Test Case 2: EAV Number Filter
+
+```typescript
+it('should filter by EAV number attribute', async () => {
+  const result = await filterService.applyFilter({
+    entity_type: 'LEAD',
+    quickFilter: [{
+      filter_attribute: 'father_sal',
+      filter_operator: 'greater_than',
+      filter_value: 100,
+      filter_entity_type: 'LFMG'
+    }],
+    loggedInUser: userData
+  });
+
+  expect(result.success).toBe(true);
+});
+```
+
+### Test Case 3: Mixed DB_COL and EAV
+
+```typescript
+it('should filter by both DB_COL and EAV attributes', async () => {
+  const result = await filterService.applyFilter({
+    entity_type: 'LEAD',
+    quickFilter: [
+      {
+        filter_attribute: 'name', // DB_COL
+        filter_operator: 'contains',
+        filter_value: 'Lead',
+        filter_entity_type: 'LEAD'
+      },
+      {
+        filter_attribute: 'father_sal', // EAV
+        filter_operator: 'equal',
+        filter_value: 100,
+        filter_entity_type: 'LFMG'
+      }
+    ],
+    loggedInUser: userData
+  });
+
+  expect(result.success).toBe(true);
+});
+```
+
+## Summary
+
+The EAV filter implementation provides:
+- ✅ Seamless filtering on both DB_COL and EAV attributes
+- ✅ Support for all standard operators (equal, contains, greater_than, etc.)
+- ✅ Sub-entity filtering with automatic relation resolution
+- ✅ Type-safe condition building for different data types
+- ✅ No breaking changes to existing API
+- ✅ Automatic EAV table detection and JOIN generation
+
+The system is production-ready and can handle complex filtering scenarios involving main entities and multiple sub-entities with mixed storage types (DB_COL and EAV).