> ## Documentation Index
> Fetch the complete documentation index at: https://docs.dataharbor.co/llms.txt
> Use this file to discover all available pages before exploring further.
# Data Transform
> Real-time field transformations applied at the data boundary.
# Data Transform
**Transform in flight, not in a pipeline.**
Real-time field transformations applied at the data boundary. No ETL pipelines or data duplication required.
Data Transform shares the same field path rules as Data Control. See [Field Targeting](./field-targeting) for nested paths, arrays, and `required` behavior.
You'll see `fields` written two ways in our examples — both are valid YAML and produce identical results:
```yaml theme={null}
# Inline style
fields: [first_name, last_name]
# Multi-line style
fields:
- first_name
- last_name
```
For a complete Virtual API Configuration, start with `version: "0.3"` and define controls under `objects..controls`. Snippets below that start with `controls:` are fragments meant to live inside an object definition.
## Operations
### Combine
Merge two or more string fields into a single new field.
```yaml theme={null}
controls:
- type: combine
fields:
- first_name
- last_name
into: full_name
separator: " "
```
**Input:**
```json theme={null}
{ "first_name": "Jane", "last_name": "Doe", "email": "jane@example.com" }
```
**Output:**
```json theme={null}
{ "first_name": "Jane", "last_name": "Doe", "full_name": "Jane Doe", "email": "jane@example.com" }
```
### Coalesce
Select the first non-empty value from two or more fields. Ideal for fallback logic and data normalization.
```yaml theme={null}
controls:
- type: coalesce
fields:
- mobile_phone
- home_phone
into: primary_phone
```
**Input:**
```json theme={null}
{ "mobile_phone": "", "home_phone": "555-1234" }
```
**Output:**
```json theme={null}
{ "mobile_phone": "", "home_phone": "555-1234", "primary_phone": "555-1234" }
```
### Delete
Remove one or more fields entirely from the outbound payload. Use this when you do not want consumers to see that a field exists at all.
```yaml theme={null}
controls:
- type: delete
fields:
- internal_note
- debug_flags
```
**Input:**
```json theme={null}
{ "full_name": "Jane Doe", "internal_note": "vip", "debug_flags": ["test"], "email": "jane@example.com" }
```
**Output:**
```json theme={null}
{ "full_name": "Jane Doe", "email": "jane@example.com" }
```
## Execution Rules
* Controls execute top-to-bottom. A later control can use a field created by an earlier control.
* `fields` order matters for both `combine` and `coalesce`.
* `combine` only operates on string source fields.
* All `fields` and `into` must share the same parent path when you use nested paths. For example, `address.city`, `address.state`, and `address.location` are valid together.
* `delete` does not use `into`. It removes the fields named in `fields` directly.
* `into` must point to a new field. If the target field already exists, the transform fails instead of overwriting it.
* Set `required: true` when missing inputs should fail the request. That check is full-path strict: if any parent segment or leaf target is missing, the request fails instead of silently skipping the control. For `coalesce`, missing, `null`, and empty-string values are treated as empty, but whitespace-only strings, `0`, and `false` still count as present values.
* `remove_source: true` works for both `combine` and `coalesce`. Use `delete` when you want an explicit cleanup step later in the control pipeline.
* If the matched payload is an array, the full control pipeline runs against each matching object element in order. Non-object elements are skipped.
* Array execution is not transactional across elements. If an earlier element has already been transformed and a later element fails a `required` check, the request still fails, but the earlier successful element changes are not rolled back.
## Nested Paths
Use nested paths when all source fields and the destination share the same parent:
```yaml theme={null}
controls:
- type: combine
fields: [address.city, address.state]
into: address.location
separator: ", "
```
## Combining with privacy controls
Controls execute top-to-bottom. If a filter control (like `hash` or `redact`) runs **before** a transform that references the same field, the transform will see the already-modified value. Place transforms before filters when both operate on the same fields.
Apply transformations before or after privacy controls:
```yaml theme={null}
controls:
# First, combine name fields
- type: combine
fields: [first_name, last_name]
into: full_name
separator: " "
# Then, anonymize the combined name
- type: anonymize
fields:
- full_name
# Redact the original fields
- type: redact
fields:
- first_name
- last_name
```
## Removing original fields
Optionally remove source fields after transformation. This works on both `combine` and `coalesce`:
```yaml theme={null}
controls:
- type: combine
fields: [first_name, last_name]
into: full_name
separator: " "
remove_source: true
```
**Output:**
```json theme={null}
{ "full_name": "Jane Doe", "email": "jane@example.com" }
```
If you prefer to keep the cleanup as a separate ordered step, use `delete` after the transform:
```yaml theme={null}
controls:
- type: combine
fields: [first_name, last_name]
into: full_name
separator: " "
- type: delete
fields: [first_name, last_name]
```
Use `delete` when the field should disappear completely. Use `redact` when the field should remain present but its value should be blanked or masked.
## Next steps
Learn how transform paths work with nested objects and arrays
Redaction, tokenization, anonymization
Detailed transform options and execution rules