Commit#
Interface#
-
class FileStoreCommit#
Interface for commit operations in a file store.
The
FileStoreCommitclass provides interfaces for committing changes, expiring old snapshots, dropping partitions, and retrieving commit metrics.Note
Direct file-system commits support append-only and primary-key tables on non-object-store paths.
CommitContextBuilder::WithCatalog()also supports object stores when the catalog manages versions.UseRESTCatalogCommit()only prepares a request for the caller to send.Public Functions
-
virtual ~FileStoreCommit() = default#
Commit changes to the file store.
- Parameters:
commit_messages – A vector of commit messages to be committed.
commit_identifier – An optional identifier for the commit operation. Default is
BATCH_WRITE_COMMIT_IDENTIFIER.watermark – An optional event-time watermark used to indicate the progress of data processing. Default is std::nullopt.
- Returns:
Status indicating the success or failure of the commit operation.
-
virtual Result<int64_t> CommitWithProgress(const std::vector<RealtimeCommitProgress> &realtime_commits, int64_t commit_identifier, std::optional<int64_t> watermark) = 0#
Commit sealed real-time segments and persist their partition-bucket offset progress.
Entries for each partition-bucket must be ordered and non-overlapping after the offset recorded by the latest committed snapshot. Input entries may be unordered; this method orders them by partition, bucket, and offset before validation. Offset gaps are allowed. The resulting snapshot atomically publishes the data files and the updated offset map.
For each partition-bucket, the upstream coordinator must include the complete prefix of prepared-but-uncommitted entries through the requested progress. Because offsets may be sparse, this method cannot distinguish a valid offset gap from an omitted prepared entry. Omitting an earlier entry may advance committed progress past unpublished files and allow their real-time data to be reclaimed.
Snapshot conflicts are retried internally using the configured commit retry limit, timeout, and backoff. Each attempt reloads the latest snapshot and rebases both file changes and offset progress. A retry succeeds idempotently when both the identifier and all requested ranges are already committed. Inconsistent identifiers, overlapping offset progress, and file or index conflicts fail without further retry.
An error is terminal for the writer state which produced
realtime_commits. The caller must discard itsRealtimeContextandFileStoreWrite, load the current latest snapshot’s durable offsets, recreate both objects, and replay input from those exclusive offsets. Conflicts fromCommitContextBuilder::WithCatalog()are retried internally. WithUseRESTCatalogCommit(), the caller sends the request and handles server conflicts. Concurrent rollback or partition deletion must be fenced by the upstream coordinator. Configure catalog writers withWriteContextBuilder::WithCatalog()too.- Parameters:
realtime_commits – Commit messages and left-closed, right-open offset ranges to commit.
commit_identifier – Identifier of the streaming commit operation.
watermark – Optional event-time watermark.
- Returns:
The id of the latest snapshot containing the committed progress. On retry, this may be a snapshot produced by a later commit and is suitable for refreshing a real-time context.
Filter out all
std::vector<CommitMessage>which have been committed and commit the remaining ones.Compared to commit, this method will first check if a commit_identifier has been committed, so this method might be slower. A common usage of this method is to retry the commit process after a failure.
- Parameters:
commit_identifier_and_messages – A map containing all
CommitMessages in question. The key is the commit_identifier.watermark – An optional event-time watermark used to indicate the progress of data processing. Default is std::nullopt.
- Returns:
Number of
std::vector<CommitMessage>committed.
Overwrite from manifest committable and partition.
Note
A full-table overwrite clears all committed real-time progress. A partition overwrite removes progress only for matching partitions. In either case, active real-time writers and their
RealtimeContextinstances must be recreated before further real-time operations.- Parameters:
partition – A single partition maps each partition key to a partition value. Depending on the user-defined statement, the partition might not include all partition keys. Also note that this partition does not necessarily equal to the partitions of the newly added key-values. This is just the partition to be cleaned up.
commit_messages – Description of the commit messages.
commit_identifier – Unique identifier.
watermark – An optional event-time watermark used to indicate the progress of data processing. Default is std::nullopt.
- Returns:
Result of the operation.
This is a temporary interface for internal use.
It will be removed in a future version. Please do not rely on it for long-term use.
Note
A full-table overwrite clears all committed real-time progress. A partition overwrite removes progress only for matching partitions. In either case, active real-time writers and their
RealtimeContextinstances must be recreated before further real-time operations.- Parameters:
partition – Description of the partition.
commit_messages – Description of the commit messages.
commit_identifier – Unique identifier.
watermark – An optional event-time watermark used to indicate the progress of data processing. Default is std::nullopt.
- Returns:
Result of the operation.
-
virtual Result<std::string> GetLastCommitTableRequest() = 0#
Returns the request from the latest commit attempt, including failed attempts.
Each attempt clears the request of the one before it, so an attempt that failed before building one, as one naming a branch a catalog cannot address does, leaves this returning an error. Catalog commits send requests automatically;
UseRESTCatalogCommit()only prepares them.Note
Temporary interface for internal use, will be removed in the future.
- Returns:
JSON with
tableId,baseSnapshotUuid,snapshotandstatistics. UnsettableIdand an absent or legacy base snapshot UUID are serialized as null.
-
virtual Result<int32_t> Expire() = 0#
Expire old snapshot in the file store.
Protects files referenced by retained snapshots, including files restored by rollback. Catalog commits require the current snapshot and retained history to be published to the file system before deletion. An unpublished or mismatched current snapshot skips expiration; catalog or retained-metadata read errors propagate without deleting files. Retained snapshots with index manifests return
NotImplementedbefore deletion.Note
Coordinate rollback and expiration so they do not run concurrently.
Note
Returns
NotImplementedbefore any snapshot is read or any file is deleted when this commit is on a branch other than main, or the table path holds such a branch underbranch/branch-<name>: branches share the table’s data files, while expiration only reads the retained snapshots of its own branch. A branch held only by a catalog, or created while expiration runs, is not found; do not expire a table with such a branch, and serialize branch creation and expiration.- Returns:
Result<int32_t> indicating the number of expired items or an error status.
-
virtual Status DropPartition(const std::vector<std::map<std::string, std::string>> &partitions, int64_t commit_identifier) = 0#
Drop specified partitions from the file store.
Note
A partition drop removes committed real-time progress only for matching partitions. Active real-time writers and their
RealtimeContextinstances must be recreated before further real-time operations.- Parameters:
partitions – A vector of partitions to be dropped.
commit_identifier – An identifier for the commit operation.
- Returns:
Status indicating the success or failure of the drop partition operation.
-
virtual Status TruncateTable(int64_t commit_identifier) = 0#
Truncate the whole table by overwriting all partitions with empty data.
The generated snapshot has commit kind OVERWRITE.
Note
Truncation clears all committed real-time progress. Active real-time writers and their
RealtimeContextinstances must be recreated before further real-time operations.- Parameters:
commit_identifier – An identifier for the commit operation.
- Returns:
Status indicating the success or failure of the truncate operation.
Abort an unsuccessful commit.
The data and index files described by the given commit messages will be deleted on a best-effort basis (delete failures are ignored).
- Parameters:
commit_messages – A vector of commit messages whose files should be cleaned up.
- Returns:
Status indicating the success or failure of the abort operation.
-
virtual Result<bool> RollbackToAsLatest(int64_t target_snapshot_id) = 0#
Roll back to the target snapshot and materialize it as the latest snapshot.
Reads the surviving files of both the current latest snapshot and the target snapshot, then commits an OVERWRITE snapshot whose visible state equals the target.
Note
Rollback restores the real-time progress recorded by the target snapshot. Active real-time writers and their
RealtimeContextinstances must be recreated before further real-time operations. Coordinate rollback with expiration so the target’s files cannot be deleted while the rollback is being prepared or committed.- Parameters:
target_snapshot_id – The snapshot id to roll back to.
- Returns:
Result<bool>; true if the atomic commit succeeded. Returns an error status if there is no latest snapshot or the target snapshot does not exist.
-
virtual FileStoreCommit &RowIdCheckConflict(std::optional<int64_t> row_id_check_from_snapshot) = 0#
Configure row-id conflict checking from a specific snapshot id.
If set to a snapshot id, commit conflict detection will additionally validate row-id conflicts against snapshots after that id. Passing std::nullopt disables this behavior.
- Parameters:
row_id_check_from_snapshot – Snapshot id to start row-id conflict checks from, or std::nullopt to disable.
- Returns:
Current commit object for chaining.
-
virtual std::shared_ptr<Metrics> GetCommitMetrics() const = 0#
Retrieve metrics related to commit operations.
- Returns:
A shared pointer to a
Metricsobject containing commit metrics.
Public Static Functions
-
static Result<std::unique_ptr<FileStoreCommit>> Create(std::unique_ptr<CommitContext> context)#
Create an instance of
FileStoreCommit.- Parameters:
context – A unique pointer to the
CommitContextused for commit operations.- Returns:
A Result containing a unique pointer to the
FileStoreCommitinstance.
-
virtual ~FileStoreCommit() = default#
-
class CommitContextBuilder#
CommitContextBuilderused to build aCommitContext, has input validation.Public Functions
-
CommitContextBuilder(const std::string &root_path, const std::string &commit_user)#
Constructs a
CommitContextBuilderwith required parameters.- Parameters:
root_path – The root path of the Paimon table.
commit_user – The user identifier for the commit operation.
Constructs a
CommitContextBuilderfor a format table that is already loaded: the only way to commit to one whose schema lives in a metastore rather than under its location, such as a table a REST catalog serves.The table carries what such a location does not say, so
WithFileSystem()is refused here rather than ignored.There is no commit user: a format table keeps no snapshot to record one in.
- Parameters:
table – The format table to commit to, as
Catalog::GetFormatTable()hands it back.
-
~CommitContextBuilder()#
-
CommitContextBuilder &SetOptions(const std::map<std::string, std::string> &options)#
Set a configuration options map to set some option entries which are not defined in the table schema or whose values you want to overwrite.
Note
The options map will clear the options added by
AddOption()before.- Parameters:
options – The configuration options map.
- Returns:
Reference to this builder for method chaining.
-
CommitContextBuilder &AddOption(const std::string &key, const std::string &value)#
Add a single configuration option which is not defined in the table schema or whose value you want to overwrite.
If you want to add multiple options, call
AddOption()multiple times or useSetOptions()instead.- Parameters:
key – The option key.
value – The option value.
- Returns:
Reference to this builder for method chaining.
-
CommitContextBuilder &IgnoreEmptyCommit(bool ignore_empty_commit)#
Sets whether to ignore empty commits (default is true).
When set to true, commits that don’t contain any actual data changes will be ignored.
- Parameters:
ignore_empty_commit – True to ignore empty commits, false otherwise.
- Returns:
Reference to this builder for method chaining.
-
CommitContextBuilder &UseRESTCatalogCommit(bool use_rest_catalog_commit)#
Builds a REST commit request for the caller to send (default is false).
Use
WithCatalog()for catalog-managed commits; the two modes are mutually exclusive.Note
Temporary interface, will be removed in the future.
- Parameters:
use_rest_catalog_commit – True to build the request without sending it.
- Returns:
Reference to this builder for method chaining.
Loads the current schema from
catalog.Catalogs with version management also load and commit snapshots, rebasing on conflicts. Other catalogs use file-system snapshot commits. Mutually exclusive with
UseRESTCatalogCommit().Data and manifests remain on the file system. Recovery, rollback and row-id checks require historical snapshots there; row-id checks also require the inspected data files’ schemas.
Expire()manages file-system snapshots after confirming that the catalog’s current snapshot and retained history are published there. Configure writers with the same catalog viaWriteContextBuilder::WithCatalog(). A branch is addressed by the identifier, astbl$branch_dev, so that the catalog answers for that branch. Its schema is read from the branch directory rather than from the catalog, just as a read of that branch reads it. Configure the writer with the same identifier.- Parameters:
catalog – Non-null catalog, kept alive by this context.
identifier – Table to commit to, naming a branch of it as
tbl$branch_dev.
- Returns:
Reference to this builder for method chaining.
-
CommitContextBuilder &WithTableId(const std::string &table_id)#
Sets the catalog table UUID sent as
tableIdto detect table recreation (default is unset).Capture
Table::CatalogUuid()before preparing changes.Table::Uuid()may fall back to the table name. An unset ID is sent as null and validated by the server. RequiresWithCatalog()orUseRESTCatalogCommit().- Parameters:
table_id – Catalog UUID of the table to commit to.
- Returns:
Reference to this builder for method chaining.
-
CommitContextBuilder &AppendCommitCheckConflict(bool append_commit_check_conflict)#
Sets whether append commits should perform conflict checking (default is false).
- Parameters:
append_commit_check_conflict – True to enable append conflict checks.
- Returns:
Reference to this builder for method chaining.
Sets the memory pool to be used for memory allocation during commit operations.
- Parameters:
memory_pool – Shared pointer to the memory pool instance.
- Returns:
Reference to this builder for method chaining.
Sets the executor to be used for asynchronous operations during commit.
- Parameters:
executor – Shared pointer to the executor instance.
- Returns:
Reference to this builder for method chaining.
Sets a custom file system instance to be used for all file operations in this commit context.
This bypasses the global file system registry and uses the provided implementation directly.
Note
If not set, use default file system (configured in
Options::FILE_SYSTEM)- Parameters:
file_system – The file system to use.
- Returns:
Reference to this builder for method chaining.
-
Result<std::unique_ptr<CommitContext>> Finish()#
Build and return a
CommitContextinstance with input validation.- Returns:
Result containing the constructed
CommitContextor an error status.
-
CommitContextBuilder(const std::string &root_path, const std::string &commit_user)#
-
class CommitContext#
CommitContextis some configuration for commit operations.Please do not use this class directly, use
CommitContextBuilderto build aCommitContextwhich has input validation.See also
Public Functions
Retained for source compatibility. Prefer
CommitContextBuilderfor input validation.
-
~CommitContext()#
-
inline const std::string &GetRootPath() const#
-
inline const std::string &GetCommitUser() const#
-
inline bool IgnoreEmptyCommit() const#
-
inline bool UseRESTCatalogCommit() const#
-
inline const std::shared_ptr<Catalog> &GetCatalog() const#
Returns the configured catalog, or null if unset.
-
inline const std::optional<Identifier> &GetIdentifier() const#
Returns the table identifier supplied with the catalog.
-
inline const std::optional<std::string> &GetTableId() const#
Returns the catalog table UUID sent as
tableIdin commit requests.
-
inline bool AppendCommitCheckConflict() const#
-
inline std::shared_ptr<MemoryPool> GetMemoryPool() const#
-
inline std::shared_ptr<FileSystem> GetSpecificFileSystem() const#
-
inline const std::map<std::string, std::string> &GetOptions() const#
-
inline const std::shared_ptr<FormatTable> &GetFormatTable() const#
The format table this context was built from, or null when it names a table path and the schema under that path says what kind of table it is.
-
class CommitMessage#
Commit message for partition and bucket.
Supports serialization and deserialization compatible with Java Paimon.
Note
Serialized payloads do not embed their serialization version. Transport
CurrentVersion()alongside the payload and pass it explicitly toDeserialize()orDeserializeList().Public Functions
-
virtual ~CommitMessage()#
Public Static Functions
-
static int32_t CurrentVersion()#
Gets the version with which this serializer serializes.
Serializes a single commit message to a binary string format.
The serialized format is compatible with Java Paimon. The serialization version is not included in the returned payload.
- Parameters:
commit_message – The commit message to serialize.
pool – Memory pool for memory allocation during serialization.
- Returns:
Result containing the serialized string data, or an error if serialization fails.
Serializes a list of commit messages to a binary string format.
The serialization version is not included in the returned payload.
- Parameters:
commit_messages – Vector of commit messages to serialize.
pool – Memory pool for memory allocation during serialization.
- Returns:
Result containing the serialized string data, or an error if serialization fails.
Deserializes a single commit message from binary data.
- Parameters:
version – The serialization format version used when the data was serialized.
buffer – Pointer to the binary data buffer.
length – Length of the binary data in bytes.
pool – Memory pool for memory allocation during deserialization.
- Returns:
Result containing the deserialized CommitMessage, or an error if deserialization fails.
Deserializes a list of commit messages from binary data.
This is the counterpart to SerializeList() for batch processing.
- Parameters:
version – The serialization format version used when the data was serialized.
buffer – Pointer to the binary data buffer.
length – Length of the binary data in bytes.
pool – Memory pool for memory allocation during deserialization.
- Returns:
Result containing a vector of deserialized CommitMessages, or an error if deserialization fails.
Converts a commit message to a human-readable debug string.
This is useful for logging, debugging, and troubleshooting purposes.
- Parameters:
commit_message – The commit message to convert to string.
- Returns:
Result containing the debug string representation, or an error if conversion fails.
-
virtual ~CommitMessage()#