Transformations

Define reusable transformations that can be attached to flows (CDC conversion, row filtering, JSON parsing, data masking, vector embedding, etc.).

Methods

Method	Description
create_transformation	Create a new transformation
delete_transformation	Delete a transformation
describe_transformation	Show full configuration for a transformation
show_transformations	List all transformations in the project

---

create_transformation

create_transformation(
    name: str,
    *,
    type: str,
    options: Mapping[str, str] | None = None,
    unsafe_raw: bool = False,
    timeout: float | None = None,
    poll_interval: float | None = None,
)

Parameter	Required	Type / values
name	yes	str
type	yes	"CONVERT_CDC", "ROW_FILTERING", "DERIVED_DATE", "ADD_CURRENT_TIMESTAMP", "USER_PROVIDED", "DATA_MASKING", "L_PARSE_JSON", "FLATTEN_STRUCT", "EXPLODE_ARRAY", "VECTOR_EMBEDDING"
options	no	Mapping[str, str] — transformation-specific config

Examples

# Row-level filter
client.create_transformation(
    "drop_test_rows",
    type="ROW_FILTERING",
    options={"filter.expression": "is_test = false"},
)

# Derived date column
client.create_transformation(
    "event_date",
    type="DERIVED_DATE",
    options={"source.field": "event_ts", "target.field": "event_date"},
)

---

delete_transformation

delete_transformation(name: str, *, unsafe_raw=False, timeout=None, poll_interval=None)

Example

client.delete_transformation("drop_test_rows")

---

describe_transformation

describe_transformation(name: str, *, unsafe_raw=False, timeout=None, poll_interval=None)

Example

result = client.describe_transformation("drop_test_rows")

---

show_transformations

show_transformations(*, timeout=None, poll_interval=None)

Example

result = client.show_transformations()