Updating Data
Every write to an Edge Shard goes through a single method, update, which takes one UpdateOperation describing what to change. The operation is written to the write-ahead log (WAL) before it is applied to storage, so with the WAL enabled, an update that has returned survives a crash. If you disable the WAL, updates are durable only after flush returns.
update
def update(self, operation: UpdateOperation) -> None
pub fn update(&self, operation: UpdateOperation) -> OperationResult<()>
| Parameter | Description |
|---|---|
operation | The operation to apply. |
Returns nothing in Python. In Rust, returns Ok(()) on success.
Creating a named vector that already exists with different parameters fails.
Update Operations
In Python, UpdateOperation is a class with static constructors, one per operation. Each returns an UpdateOperation to pass to update.
| Operation | Parameters | Description |
|---|---|---|
upsert_points | points, condition, update_mode | Insert or update points. |
delete_points | point_ids | Delete points by ID. |
delete_points_by_filter | filter | Delete every point matching a filter. |
update_vectors | point_vectors, condition | Replace vectors on existing points. |
delete_vectors | point_ids, vector_names | Delete named vectors from points. |
delete_vectors_by_filter | filter, vector_names | Delete named vectors from matching points. |
set_payload | point_ids, payload, key | Merge payload fields into points. |
set_payload_by_filter | filter, payload, key | Merge payload fields into matching points. |
overwrite_payload | point_ids, payload, key | Replace the entire payload on points. |
overwrite_payload_by_filter | filter, payload, key | Replace the entire payload on matching points. |
delete_payload | point_ids, keys | Delete named payload fields from points. |
delete_payload_by_filter | filter, keys | Delete named payload fields from matching points. |
clear_payload | point_ids | Delete all payload from points. |
clear_payload_by_filter | filter | Delete all payload from matching points. |
Payload Indexes
The update operation also enables you to create and delete payload indexes.
| Operation | Parameters | Description |
|---|---|---|
create_field_index | field_name, schema | Index a payload field. |
delete_field_index | field_name | Remove a payload field index. |
The schema of create_field_index takes the index parameters for the field type, for example KeywordIndexParams or TextIndexParams. Every index parameters type accepts:
| Parameter | Description |
|---|---|
memory | Memory tier of the index. |
nested | Index each object of an array as a separate entry. field_name must be a path inside the array, for example offers[].price with nested set to "offers". An is_empty condition on the field matches a point when one of the array’s elements lacks the field. |
TextIndexParams also takes scoring, which records document lengths and term frequencies so that a TextQuery can rank points over the field with BM25. Refer to BM25 with Qdrant Edge.
Modify the Vector Schema
Add or remove named vectors to an existing Edge Shard’s schema. This is useful when migrating to a new embedding model or adding hybrid search to an Edge Shard that already contains data.
| Operation | Parameters | Description |
|---|---|---|
create_dense_vector | vector_name, size, distance, multivector_config, datatype | Add a dense named vector to the schema. |
create_sparse_vector | vector_name, modifier, datatype | Add a sparse named vector to the schema. |
delete_vector_name | vector_name | Remove a named vector from the schema. |
Parameters
The key parameter on the payload operations targets a nested field path rather than the payload root.
The condition parameter gates the write by a filter, but the two operations differ. On update_vectors it applies to the whole operation: only points matching the filter are updated. On upsert_points it applies only to points that already exist, so an existing point that fails the filter is left untouched while a point that is new to the shard is inserted regardless of the filter.
The condition parameter controls whether a write is applied, but its behavior depends on the operation.
For update_vectors, only points that match the filter are updated. For upsert_points, the filter applies only to existing points. Existing points that do not match the filter remain unchanged, while new points are inserted regardless of the filter.
upsert_points also takes an update_mode:
| Mode | Behavior |
|---|---|
UpdateMode.Upsert | Insert new points and update existing ones. This is the default behavior. |
UpdateMode.InsertOnly | Insert new points, leave existing ones untouched. |
UpdateMode.UpdateOnly | Update existing points, do not insert new ones. |
In Rust, UpdateOperation is an enum that groups operations by what they modify:
| Variant | Covers |
|---|---|
PointOperation | Upserting, deleting, and syncing points. |
VectorOperation | Updating and deleting named vectors on existing points. |
PayloadOperation | Setting, overwriting, deleting, and clearing payload. |
FieldIndexOperation | Creating and deleting payload field indexes. |
VectorNameOperation | Adding and removing named vectors from the schema. |
Ready to try Qdrant?