Transaction

CommitProperties and PostCommitHookProperties

deltalake.CommitProperties dataclass

CommitProperties(custom_metadata: dict[str, Any] | None = None, max_commit_retries: int | None = None, app_transactions: list[Transaction] | None = None)

The commit properties. Controls the behaviour of the commit.

Custom metadata to be stored in the commit. Controls the number of retries for the commit.

Parameters:

Name	Type	Description	Default
custom_metadata	dict[str, Any] | None	custom metadata that will be added to the transaction commit. Top level keys must be strings and values must be JSON serializable. Values such as NaN and Infinity are rejected. Reserved commit info keys have stricter requirements: "operationParameters" must be a JSON object; "readVersion" must be an integer greater than or equal to zero, not a float; "userId", "userName", and "userMetadata" must be strings; "isolationLevel" must be one of "Serializable", "WriteSerializable", or "SnapshotIsolation"; "isBlindAppend" must be a boolean. The generated keys "timestamp", "operation", and "engineInfo" cannot be set through custom metadata. Callers may set "clientVersion", and it is preserved when provided.	None
max_commit_retries	int | None	maximum number of times to retry the transaction commit.	None

deltalake.PostCommitHookProperties dataclass

PostCommitHookProperties(create_checkpoint: bool = True, cleanup_expired_logs: bool | None = None)

The post commit hook properties, only required for advanced usecases where you need to control this.

Checkpoints are by default created based on the delta.checkpointInterval config setting. cleanup_expired_logs can be set to override the delta.enableExpiredLogCleanup, otherwise the config setting will be used to decide whether to clean up logs automatically by taking also the delta.logRetentionDuration into account.

Parameters:

Name	Type	Description	Default
create_checkpoint	bool	to create checkpoints based on checkpoint interval. Defaults to True.	True
cleanup_expired_logs	Optional[bool]	to clean up logs based on interval. Defaults to None.	None

Create write transactions

deltalake.transaction.AddAction dataclass

AddAction(path: str, size: int, partition_values: Mapping[str, str | None], modification_time: int, data_change: bool, stats: str)

deltalake.transaction.RemoveAction

RemoveAction(path: str, data_change: bool, deletion_timestamp: int, size: int | None = None, partition_values: dict[str, str | None] | None = None)

deltalake.transaction.create_table_with_add_actions

create_table_with_add_actions(table_uri: str, schema: Schema, add_actions: list[AddAction], mode: Literal['error', 'append', 'overwrite', 'ignore'] = 'error', partition_by: list[str] | str | None = None, name: str | None = None, description: str | None = None, configuration: Mapping[str, str | None] | None = None, storage_options: dict[str, str] | None = None, *args: Any, commit_properties: CommitProperties | None = None, post_commithook_properties: PostCommitHookProperties | None = None) -> None

deltalake.DeltaTable.create_write_transaction

create_write_transaction(actions: list[AddAction | RemoveAction], mode: str, schema: DeltaSchema | ArrowSchemaExportable, partition_by: list[str] | str | None = None, partition_filters: FilterType | None = None, *args: Any, commit_properties: CommitProperties | None = None, post_commithook_properties: PostCommitHookProperties | None = None) -> None

Commit file actions to the table as a single transaction.

Files referenced by an AddAction must already exist in the table location. A RemoveAction marks an existing file as removed and is only allowed with mode="append"; mode="overwrite" derives its own remove actions from the partition filters. A path cannot be both added and removed in the same transaction, and tables with delta.appendOnly=true reject remove actions that have data_change=True.

Parameters:

Name	Type	Description	Default
actions	list[AddAction | RemoveAction]	AddAction and RemoveAction entries to commit together.	required
mode	str	save mode of the write, "append" or "overwrite".	required
schema	Schema | ArrowSchemaExportable	schema of the files being added.	required
partition_by	list[str] | str | None	partition columns of the table.	None
partition_filters	FilterType | None	partitions to replace when mode="overwrite".	None
commit_properties	CommitProperties | None	properties of the transaction commit.	None
post_commithook_properties	PostCommitHookProperties | None	properties for the post commit hook.	None