Transformation Engine

The transformation engine maps model properties to target representations through nine transformation types, an expression builder, and automatic conflict detection.

April 1, 2026

Nine Transformation Types

The transformation engine provides nine distinct mapping types. Each serves a specific purpose, from straightforward field copies to complex multi-field calculations. Choosing the right type makes the intent of each mapping explicit and traceable.

Type	Description	Example
Direct	Copy the value unchanged from source to target.	Model <em>email</em> maps to target <em>email</em> with the same type.
Rename	Copy the value but change the field name to match target conventions.	Model <em>firstName</em> becomes <em>first_name</em> in PostgreSQL.
Type-Cast	Convert the value to a different data type.	A string ISO date becomes a native <em>timestamptz</em> column.
Aggregate	Compute an aggregate over related entities.	Count of <em>Order</em> entities becomes <em>order_count</em> on <em>Customer</em>.
Concatenate	Combine multiple source fields into one target field.	<em>firstName</em> and <em>lastName</em> join to form <em>full_name</em>.
Calculate	Apply arithmetic or date logic to produce a derived value.	<em>dateOfBirth</em> becomes <em>age</em> via date subtraction.
Derive	Create a field with no direct source in the model.	A surrogate <em>id</em> column or an <em>updated_at</em> audit timestamp.
Split	Break one source field into multiple target fields.	An <em>address</em> string splits into <em>street</em>, <em>city</em>, <em>postcode</em>.
Custom	Write a free-form expression for cases that do not fit the other types.	CASE/WHEN logic that maps status codes to human-readable labels.

Expression Builder

The expression builder provides a visual interface for constructing transformation logic. Rather than writing raw SQL or code, you compose expressions from field references, functions, and operators. The builder validates expressions as you type and shows the resolved output type so you can catch mismatches early.

Automatic Conflict Detection

When source and target types do not align, the transformation engine detects the mismatch and categorises it by severity. Conflicts appear inline in the mapper so you can resolve them before saving. This prevents silent data loss that would otherwise surface only at deployment time.

Type Conversion

The source type does not match the target type. The engine suggests appropriate CAST operations and shows whether the conversion is lossless or potentially lossy.

Truncation

The target field has a smaller capacity than the source. A varchar(50) target for a varchar(255) source flags the risk of data being cut off.

Null Handling

The source allows null values but the target does not, or vice versa. The engine flags the mismatch and suggests default value or coalesce strategies.

Precision Loss

Numeric conversions that reduce precision, such as decimal to integer or double to float. The engine quantifies the potential loss.

Bulk Operations

When working with entities that have many properties, individual mapping becomes tedious. The transformation engine supports batch operations: auto-map properties by name similarity, apply a transformation type to all selected properties at once, and bulk-rename using pattern rules (such as camelCase to snake_case). These operations respect existing mappings and only affect unmapped or selected properties.