Many resources are needed to download a project. Please understand that we have to compensate our server costs. Thank you in advance. Project price only 1 $
You can buy this project and download/modify it how often you want.
table = getTable(databaseName, tableName);
if (table.isEmpty()) {
return Optional.empty();
}
List partitionNames;
TableSource tableSource = getTableSource(databaseName, tableName);
partitionNames = switch (tableSource) {
case CREATED_IN_THIS_TRANSACTION -> ImmutableList.of();
case PRE_EXISTING_TABLE -> getOptionalPartitions(databaseName, tableName, columnNames, partitionKeysFilter)
.orElseThrow(() -> new TrinoException(TRANSACTION_CONFLICT, format("Table '%s.%s' was dropped by another transaction", databaseName, tableName)));
};
Set duplicatePartitionNames = ImmutableMultiset.copyOf(partitionNames)
.entrySet().stream()
.filter(entry -> entry.getCount() > 1)
.map(Multiset.Entry::getElement)
.collect(toImmutableSet());
if (!duplicatePartitionNames.isEmpty()) {
throw new TrinoException(HIVE_METASTORE_ERROR, format("Metastore returned duplicate partition names for %s", duplicatePartitionNames));
}
Map, Action> partitionActionsOfTable = partitionActions.computeIfAbsent(table.get().getSchemaTableName(), k -> new HashMap<>());
ImmutableList.Builder resultBuilder = ImmutableList.builder();
// alter/remove newly altered/dropped partitions from the results from underlying metastore
for (String partitionName : partitionNames) {
List partitionValues = toPartitionValues(partitionName);
Action partitionAction = partitionActionsOfTable.get(partitionValues);
if (partitionAction == null) {
resultBuilder.add(partitionName);
continue;
}
switch (partitionAction.type()) {
case ADD -> throw new TrinoException(TRANSACTION_CONFLICT, format("Another transaction created partition %s in table %s.%s", partitionValues, databaseName, tableName));
case DROP, DROP_PRESERVE_DATA -> {
// do nothing
}
case ALTER, INSERT_EXISTING, MERGE -> resultBuilder.add(partitionName);
}
}
// add newly added partitions to the results from underlying metastore.
if (!partitionActionsOfTable.isEmpty()) {
for (Action partitionAction : partitionActionsOfTable.values()) {
if (partitionAction.type() == ActionType.ADD) {
List values = partitionAction.data().partition().getValues();
resultBuilder.add(makePartName(columnNames, values));
}
}
}
return Optional.of(resultBuilder.build());
}
public synchronized Map> getPartitionsByNames(String databaseName, String tableName, List partitionNames)
{
checkReadable();
TableSource tableSource = getTableSource(databaseName, tableName);
Map, Action> partitionActionsOfTable = partitionActions.computeIfAbsent(new SchemaTableName(databaseName, tableName), k -> new HashMap<>());
ImmutableList.Builder partitionNamesToQueryBuilder = ImmutableList.builder();
ImmutableMap.Builder> resultBuilder = ImmutableMap.builder();
for (String partitionName : partitionNames) {
List partitionValues = toPartitionValues(partitionName);
Action partitionAction = partitionActionsOfTable.get(partitionValues);
if (partitionAction == null) {
switch (tableSource) {
case PRE_EXISTING_TABLE -> partitionNamesToQueryBuilder.add(partitionName);
case CREATED_IN_THIS_TRANSACTION -> resultBuilder.put(partitionName, Optional.empty());
}
}
else {
resultBuilder.put(partitionName, getPartitionFromPartitionAction(partitionAction));
}
}
List partitionNamesToQuery = partitionNamesToQueryBuilder.build();
if (!partitionNamesToQuery.isEmpty()) {
Map> delegateResult = getOptionalPartitions(
databaseName,
tableName,
partitionNamesToQuery);
resultBuilder.putAll(delegateResult);
}
return resultBuilder.buildOrThrow();
}
private static Optional getPartitionFromPartitionAction(Action partitionAction)
{
return switch (partitionAction.type()) {
case ADD, ALTER, INSERT_EXISTING, MERGE -> Optional.of(partitionAction.data().getAugmentedPartitionForInTransactionRead());
case DROP, DROP_PRESERVE_DATA -> Optional.empty();
};
}
public synchronized void addPartition(
ConnectorSession session,
String databaseName,
String tableName,
Partition partition,
Location currentLocation,
Optional> files,
PartitionStatistics statistics,
boolean cleanExtraOutputFilesOnCommit)
{
setShared();
checkArgument(getQueryId(partition).isPresent());
Map, Action> partitionActionsOfTable = partitionActions.computeIfAbsent(new SchemaTableName(databaseName, tableName), k -> new HashMap<>());
Action oldPartitionAction = partitionActionsOfTable.get(partition.getValues());
if (oldPartitionAction == null) {
partitionActionsOfTable.put(
partition.getValues(),
new Action<>(ActionType.ADD, new PartitionAndMore(partition, currentLocation, files, statistics, statistics, cleanExtraOutputFilesOnCommit), session.getIdentity(), session.getQueryId()));
return;
}
switch (oldPartitionAction.type()) {
case DROP, DROP_PRESERVE_DATA -> {
if (!oldPartitionAction.identity().getUser().equals(session.getUser())) {
throw new TrinoException(TRANSACTION_CONFLICT, "Operation on the same partition with different user in the same transaction is not supported");
}
partitionActionsOfTable.put(
partition.getValues(),
new Action<>(ActionType.ALTER, new PartitionAndMore(partition, currentLocation, files, statistics, statistics, cleanExtraOutputFilesOnCommit), session.getIdentity(), session.getQueryId()));
}
case ADD, ALTER, INSERT_EXISTING, MERGE ->
throw new TrinoException(ALREADY_EXISTS, format("Partition already exists for table '%s.%s': %s", databaseName, tableName, partition.getValues()));
}
}
public synchronized void dropPartition(ConnectorSession session, String databaseName, String tableName, List partitionValues, boolean deleteData)
{
setShared();
Map, Action> partitionActionsOfTable = partitionActions.computeIfAbsent(new SchemaTableName(databaseName, tableName), k -> new HashMap<>());
Action oldPartitionAction = partitionActionsOfTable.get(partitionValues);
if (oldPartitionAction == null) {
if (deleteData) {
partitionActionsOfTable.put(partitionValues, new Action<>(ActionType.DROP, null, session.getIdentity(), session.getQueryId()));
}
else {
partitionActionsOfTable.put(partitionValues, new Action<>(ActionType.DROP_PRESERVE_DATA, null, session.getIdentity(), session.getQueryId()));
}
return;
}
switch (oldPartitionAction.type()) {
case DROP, DROP_PRESERVE_DATA -> throw new PartitionNotFoundException(new SchemaTableName(databaseName, tableName), partitionValues);
case ADD, ALTER, INSERT_EXISTING, MERGE ->
throw new TrinoException(NOT_SUPPORTED, "dropping a partition added in the same transaction is not supported: %s %s %s"
.formatted(databaseName, tableName, partitionValues));
}
}
public synchronized void finishInsertIntoExistingPartitions(
ConnectorSession session,
String databaseName,
String tableName,
List partitionUpdateInfos,
boolean cleanExtraOutputFilesOnCommit)
{
setShared();
Table table = getExistingTable(databaseName, tableName);
Map, Action> partitionActionsOfTable = partitionActions.computeIfAbsent(table.getSchemaTableName(), k -> new HashMap<>());
for (PartitionUpdateInfo partitionInfo : partitionUpdateInfos) {
Action oldPartitionAction = partitionActionsOfTable.get(partitionInfo.partitionValues());
if (oldPartitionAction != null) {
switch (oldPartitionAction.type()) {
case DROP, DROP_PRESERVE_DATA -> throw new PartitionNotFoundException(table.getSchemaTableName(), partitionInfo.partitionValues());
case ADD, ALTER, INSERT_EXISTING, MERGE -> throw new UnsupportedOperationException("Inserting into a partition that were added, altered, or inserted into in the same transaction is not supported");
default -> throw new IllegalStateException("Unknown action type: " + oldPartitionAction.type());
}
}
}
// new data will on include current table columns
// partition column stats do not include the partition keys
Set columnNames = table.getDataColumns().stream()
.map(Column::getName)
.collect(toImmutableSet());
for (List partitionInfoBatch : Iterables.partition(partitionUpdateInfos, 100)) {
List partitionNames = partitionInfoBatch.stream()
.map(PartitionUpdateInfo::partitionValues)
.map(partitionValues -> getPartitionName(databaseName, tableName, partitionValues))
.collect(toImmutableList());
Map partitionsByNames = getExistingPartitions(databaseName, tableName, partitionNames);
Map basicStats = partitionsByNames.entrySet().stream()
.collect(toImmutableMap(Entry::getKey, entry -> getHiveBasicStatistics(entry.getValue().getParameters())));
Map> columnStats = delegate.getPartitionColumnStatistics(databaseName, tableName, basicStats.keySet(), columnNames);
for (int i = 0; i < partitionInfoBatch.size(); i++) {
PartitionUpdateInfo partitionInfo = partitionInfoBatch.get(i);
String partitionName = partitionNames.get(i);
Partition partition = partitionsByNames.get(partitionName);
PartitionStatistics currentStatistics = new PartitionStatistics(basicStats.get(partitionName), columnStats.get(partitionName));
partitionActionsOfTable.put(
partitionInfo.partitionValues(),
new Action<>(
ActionType.INSERT_EXISTING,
new PartitionAndMore(
partition,
partitionInfo.currentLocation(),
Optional.of(partitionInfo.fileNames()),
MERGE_INCREMENTAL.updatePartitionStatistics(currentStatistics, partitionInfo.statisticsUpdate()),
partitionInfo.statisticsUpdate(),
cleanExtraOutputFilesOnCommit),
session.getIdentity(),
session.getQueryId()));
}
}
}
private synchronized AcidTransaction getRequiredAcidTransaction()
{
return currentHiveTransaction.orElseThrow(() -> new IllegalStateException("currentHiveTransaction not present")).getTransaction();
}
private synchronized AcidTransaction getOptionalAcidTransaction()
{
return currentHiveTransaction.map(HiveTransaction::getTransaction).orElse(NO_ACID_TRANSACTION);
}
private String getPartitionName(String databaseName, String tableName, List partitionValues)
{
Table table = getTable(databaseName, tableName)
.orElseThrow(() -> new TableNotFoundException(new SchemaTableName(databaseName, tableName)));
return getPartitionName(table, partitionValues);
}
private static String getPartitionName(Table table, List partitionValues)
{
List columnNames = table.getPartitionColumns().stream()
.map(Column::getName)
.collect(toImmutableList());
return makePartName(columnNames, partitionValues);
}
@Override
public synchronized void createRole(String role, String grantor)
{
setExclusive(delegate -> delegate.createRole(role, grantor));
}
@Override
public synchronized void dropRole(String role)
{
setExclusive(delegate -> delegate.dropRole(role));
}
@Override
public synchronized Set listRoles()
{
checkReadable();
return delegate.listRoles();
}
@Override
public synchronized void grantRoles(Set roles, Set grantees, boolean adminOption, HivePrincipal grantor)
{
setExclusive(delegate -> delegate.grantRoles(roles, grantees, adminOption, grantor));
}
@Override
public synchronized void revokeRoles(Set roles, Set grantees, boolean adminOption, HivePrincipal grantor)
{
setExclusive(delegate -> delegate.revokeRoles(roles, grantees, adminOption, grantor));
}
@Override
public synchronized Set listRoleGrants(HivePrincipal principal)
{
checkReadable();
return delegate.listRoleGrants(principal);
}
@Override
public Optional getDatabaseOwner(String databaseName)
{
Database database = getDatabase(databaseName)
.orElseThrow(() -> new SchemaNotFoundException(databaseName));
return database.getOwnerName().map(ownerName -> new HivePrincipal(database.getOwnerType().orElseThrow(), ownerName));
}
@Override
public synchronized Set listTablePrivileges(String databaseName, String tableName, Optional principal)
{
checkReadable();
SchemaTableName schemaTableName = new SchemaTableName(databaseName, tableName);
Action tableAction = tableActions.get(schemaTableName);
if (tableAction == null) {
return delegate.listTablePrivileges(databaseName, tableName, getExistingTable(databaseName, tableName).getOwner(), principal);
}
return switch (tableAction.type()) {
case ADD, ALTER -> {
if (principal.isPresent() && principal.get().getType() == PrincipalType.ROLE) {
yield ImmutableSet.of();
}
Optional owner = tableAction.data().getTable().getOwner();
if (owner.isEmpty()) {
// todo the existing logic below seem off. Only permissions held by the table owner are returned
yield ImmutableSet.of();
}
String ownerUsername = owner.orElseThrow();
if (principal.isPresent() && !principal.get().getName().equals(ownerUsername)) {
yield ImmutableSet.of();
}
Collection privileges = tableAction.data().getPrincipalPrivileges().getUserPrivileges().get(ownerUsername);
yield ImmutableSet.builder()
.addAll(privileges)
.add(new HivePrivilegeInfo(OWNERSHIP, true, new HivePrincipal(USER, ownerUsername), new HivePrincipal(USER, ownerUsername)))
.build();
}
case INSERT_EXISTING, MERGE -> delegate.listTablePrivileges(databaseName, tableName, getExistingTable(databaseName, tableName).getOwner(), principal);
case DROP -> throw new TableNotFoundException(schemaTableName);
case DROP_PRESERVE_DATA -> throw new IllegalStateException("Unsupported action type: " + tableAction.type());
};
}
private synchronized String getRequiredTableOwner(String databaseName, String tableName)
{
return getExistingTable(databaseName, tableName).getOwner().orElseThrow();
}
private Table getExistingTable(String databaseName, String tableName)
{
return delegate.getTable(databaseName, tableName)
.orElseThrow(() -> new TableNotFoundException(new SchemaTableName(databaseName, tableName)));
}
private Map getExistingPartitions(String databaseName, String tableName, Collection partitionNames)
{
return getOptionalPartitions(databaseName, tableName, ImmutableList.copyOf(partitionNames)).entrySet().stream()
.collect(toImmutableMap(Entry::getKey, entry -> entry.getValue()
.orElseThrow(() -> new PartitionNotFoundException(new SchemaTableName(databaseName, tableName), extractPartitionValues(entry.getKey())))));
}
private Map> getOptionalPartitions(String databaseName, String tableName, List partitionNames)
{
return delegate.getTable(databaseName, tableName)
.map(table -> getOptionalPartitions(table, partitionNames))
.orElseGet(() -> partitionNames.stream()
.collect(toImmutableMap(name -> name, _ -> Optional.empty())));
}
private Map> getOptionalPartitions(Table table, List partitionNames)
{
if (partitionProjectionEnabled) {
Optional projection = getPartitionProjectionFromTable(table, typeManager);
if (projection.isPresent()) {
return projection.get().getProjectedPartitionsByNames(table, partitionNames);
}
}
return delegate.getPartitionsByNames(table, partitionNames);
}
private Optional> getOptionalPartitions(
String databaseName,
String tableName,
List columnNames,
TupleDomain partitionKeysFilter)
{
if (partitionProjectionEnabled) {
Table table = getTable(databaseName, tableName)
.orElseThrow(() -> new TrinoException(HIVE_TABLE_DROPPED_DURING_QUERY, "Table does not exists: " + tableName));
Optional projection = getPartitionProjectionFromTable(table, typeManager);
if (projection.isPresent()) {
return projection.get().getProjectedPartitionNamesByFilter(columnNames, partitionKeysFilter);
}
}
return delegate.getPartitionNamesByFilter(databaseName, tableName, columnNames, partitionKeysFilter);
}
@Override
public synchronized void grantTablePrivileges(String databaseName, String tableName, HivePrincipal grantee, HivePrincipal grantor, Set privileges, boolean grantOption)
{
setExclusive(delegate -> delegate.grantTablePrivileges(databaseName, tableName, getRequiredTableOwner(databaseName, tableName), grantee, grantor, privileges, grantOption));
}
@Override
public synchronized void revokeTablePrivileges(String databaseName, String tableName, HivePrincipal grantee, HivePrincipal grantor, Set privileges, boolean grantOption)
{
setExclusive(delegate -> delegate.revokeTablePrivileges(databaseName, tableName, getRequiredTableOwner(databaseName, tableName), grantee, grantor, privileges, grantOption));
}
public synchronized boolean functionExists(SchemaFunctionName name, String signatureToken)
{
checkReadable();
return delegate.functionExists(name.getSchemaName(), name.getFunctionName(), signatureToken);
}
public synchronized Collection getFunctions(String schemaName)
{
checkReadable();
return delegate.getAllFunctions(schemaName);
}
public synchronized Collection getFunctions(SchemaFunctionName name)
{
checkReadable();
return delegate.getFunctions(name.getSchemaName(), name.getFunctionName());
}
public synchronized void createFunction(SchemaFunctionName name, LanguageFunction function)
{
setExclusive(delegate -> delegate.createFunction(name.getSchemaName(), name.getFunctionName(), function));
}
public synchronized void replaceFunction(SchemaFunctionName name, LanguageFunction function)
{
setExclusive(delegate -> delegate.replaceFunction(name.getSchemaName(), name.getFunctionName(), function));
}
public synchronized void dropFunction(SchemaFunctionName name, String signatureToken)
{
setExclusive(delegate -> delegate.dropFunction(name.getSchemaName(), name.getFunctionName(), signatureToken));
}
public synchronized String declareIntentionToWrite(ConnectorSession session, WriteMode writeMode, Location stagingPathRoot, SchemaTableName schemaTableName)
{
setShared();
if (writeMode == WriteMode.DIRECT_TO_TARGET_EXISTING_DIRECTORY) {
Map, Action> partitionActionsOfTable = partitionActions.get(schemaTableName);
if (partitionActionsOfTable != null && !partitionActionsOfTable.isEmpty()) {
throw new TrinoException(NOT_SUPPORTED, "Cannot insert into a table with a partition that has been modified in the same transaction when Trino is configured to skip temporary directories.");
}
}
ConnectorIdentity identity = session.getIdentity();
String queryId = session.getQueryId();
String declarationId = queryId + "_" + declaredIntentionsToWriteCounter;
declaredIntentionsToWriteCounter++;
declaredIntentionsToWrite.add(new DeclaredIntentionToWrite(declarationId, writeMode, identity, queryId, stagingPathRoot, schemaTableName));
return declarationId;
}
public synchronized void dropDeclaredIntentionToWrite(String declarationId)
{
boolean removed = declaredIntentionsToWrite.removeIf(intention -> intention.declarationId().equals(declarationId));
if (!removed) {
throw new IllegalArgumentException("Declaration with id " + declarationId + " not found");
}
}
public synchronized boolean isFinished()
{
return state == State.FINISHED;
}
public synchronized void commit()
{
try {
switch (state) {
case EMPTY -> {}
case SHARED_OPERATION_BUFFERED -> commitShared();
case EXCLUSIVE_OPERATION_BUFFERED -> bufferedExclusiveOperation.execute(delegate);
case FINISHED -> throw new IllegalStateException("Tried to commit buffered metastore operations after transaction has been committed/aborted");
}
}
finally {
state = State.FINISHED;
}
}
public synchronized void rollback()
{
try {
switch (state) {
case EMPTY, EXCLUSIVE_OPERATION_BUFFERED -> {}
case SHARED_OPERATION_BUFFERED -> rollbackShared();
case FINISHED -> throw new IllegalStateException("Tried to rollback buffered metastore operations after transaction has been committed/aborted");
}
}
finally {
state = State.FINISHED;
}
}
public void checkSupportsHiveAcidTransactions()
{
delegate.checkSupportsTransactions();
}
public void beginQuery(ConnectorSession session)
{
String queryId = session.getQueryId();
synchronized (this) {
checkState(
currentQueryId.isEmpty() && hiveTransactionSupplier.isEmpty(),
"Query already begun: %s while starting query %s",
currentQueryId,
queryId);
currentQueryId = Optional.of(queryId);
hiveTransactionSupplier = Optional.of(() -> makeHiveTransaction(session, transactionId -> NO_ACID_TRANSACTION));
}
}
public AcidTransaction beginInsert(ConnectorSession session, Table table)
{
return beginOperation(session, table, AcidOperation.INSERT);
}
public AcidTransaction beginMerge(ConnectorSession session, Table table)
{
return beginOperation(session, table, AcidOperation.MERGE);
}
private AcidTransaction beginOperation(ConnectorSession session, Table table, AcidOperation operation)
{
String queryId = session.getQueryId();
synchronized (this) {
currentQueryId = Optional.of(queryId);
// We start the transaction immediately, and allocate the write lock and the writeId,
// because we need the writeId to write the delta files.
HiveTransaction hiveTransaction = makeHiveTransaction(session, transactionId -> {
acquireTableWriteLock(
new AcidTransactionOwner(session.getUser()),
queryId,
transactionId,
table.getDatabaseName(),
table.getTableName(),
operation,
!table.getPartitionColumns().isEmpty());
long writeId = allocateWriteId(table.getDatabaseName(), table.getTableName(), transactionId);
return new AcidTransaction(operation, transactionId, writeId);
});
hiveTransactionSupplier = Optional.of(() -> hiveTransaction);
currentHiveTransaction = Optional.of(hiveTransaction);
return hiveTransaction.getTransaction();
}
}
private HiveTransaction makeHiveTransaction(ConnectorSession session, Function transactionMaker)
{
String queryId = session.getQueryId();
long heartbeatInterval = configuredTransactionHeartbeatInterval
.map(Duration::toMillis)
.orElseGet(this::getServerExpectedHeartbeatIntervalMillis);
// TODO consider adding query id to the owner
long transactionId = delegate.openTransaction(new AcidTransactionOwner(session.getUser()));
log.debug("Using hive transaction %s for %s", transactionId, queryId);
ScheduledFuture heartbeatTask = heartbeatExecutor.scheduleAtFixedRate(
() -> delegate.sendTransactionHeartbeat(transactionId),
0,
heartbeatInterval,
MILLISECONDS);
AcidTransaction transaction = transactionMaker.apply(transactionId);
return new HiveTransaction(queryId, transactionId, heartbeatTask, transaction);
}
private long getServerExpectedHeartbeatIntervalMillis()
{
String timeout = delegate.getConfigValue("metastore.txn.timeout").orElse("300s");
return metastoreTimeToMillis(timeout) / 2;
}
private static final Pattern METASTORE_TIME = Pattern.compile("([0-9]+)([a-zA-Z]+)");
// based on org.apache.hadoop.hive.metastore.conf.MetastoreConf#convertTimeStr
private static long metastoreTimeToMillis(String value)
{
if (CharMatcher.inRange('0', '9').matches(value.charAt(value.length() - 1))) {
return SECONDS.toMillis(parseLong(value));
}
Matcher matcher = METASTORE_TIME.matcher(value);
checkArgument(matcher.matches(), "Invalid time unit: %s", value);
long duration = parseLong(matcher.group(1));
String unit = matcher.group(2).toLowerCase(ENGLISH);
if (unit.equals("s") || unit.startsWith("sec")) {
return SECONDS.toMillis(duration);
}
if (unit.equals("ms") || unit.startsWith("msec")) {
return duration;
}
if (unit.equals("m") || unit.startsWith("min")) {
return MINUTES.toMillis(duration);
}
if (unit.equals("us") || unit.startsWith("usec")) {
return MICROSECONDS.toMillis(duration);
}
if (unit.equals("ns") || unit.startsWith("nsec")) {
return NANOSECONDS.toMillis(duration);
}
if (unit.equals("h") || unit.startsWith("hour")) {
return HOURS.toMillis(duration);
}
if (unit.equals("d") || unit.startsWith("day")) {
return DAYS.toMillis(duration);
}
throw new IllegalArgumentException("Invalid time unit " + unit);
}
public Optional getValidWriteIds(ConnectorSession session, HiveTableHandle tableHandle)
{
HiveTransaction hiveTransaction;
synchronized (this) {
String queryId = session.getQueryId();
checkState(currentQueryId.equals(Optional.of(queryId)), "Invalid query id %s while current query is %s", queryId, currentQueryId);
if (!isTransactionalTable(tableHandle.getTableParameters().orElseThrow(() -> new IllegalStateException("tableParameters missing")))) {
return Optional.empty();
}
if (currentHiveTransaction.isEmpty()) {
currentHiveTransaction = Optional.of(hiveTransactionSupplier
.orElseThrow(() -> new IllegalStateException("hiveTransactionSupplier is not set"))
.get());
}
hiveTransaction = currentHiveTransaction.get();
}
return Optional.of(hiveTransaction.getValidWriteIds(new AcidTransactionOwner(session.getUser()), delegate, tableHandle));
}
public synchronized void cleanupQuery(ConnectorSession session)
{
String queryId = session.getQueryId();
checkState(currentQueryId.equals(Optional.of(queryId)), "Invalid query id %s while current query is %s", queryId, currentQueryId);
Optional transaction = currentHiveTransaction;
if (transaction.isEmpty()) {
clearCurrentTransaction();
return;
}
try {
commit();
}
catch (Throwable commitFailure) {
try {
postCommitCleanup(transaction, false);
}
catch (Throwable cleanupFailure) {
if (cleanupFailure != commitFailure) {
commitFailure.addSuppressed(cleanupFailure);
}
}
throw commitFailure;
}
postCommitCleanup(transaction, true);
}
private void postCommitCleanup(Optional transaction, boolean commit)
{
clearCurrentTransaction();
long transactionId = transaction.orElseThrow().getTransactionId();
ScheduledFuture heartbeatTask = transaction.get().getHeartbeatTask();
heartbeatTask.cancel(true);
if (commit) {
// Any failure around aborted transactions, etc. would be handled by Hive Metastore commit, and TrinoException will be thrown
delegate.commitTransaction(transactionId);
}
else {
delegate.abortTransaction(transactionId);
}
}
private synchronized void clearCurrentTransaction()
{
currentQueryId = Optional.empty();
currentHiveTransaction = Optional.empty();
hiveTransactionSupplier = Optional.empty();
}
@GuardedBy("this")
private void commitShared()
{
checkHoldsLock();
AcidTransaction transaction = getOptionalAcidTransaction();
Committer committer = new Committer(transaction);
try {
for (Entry> entry : tableActions.entrySet()) {
SchemaTableName schemaTableName = entry.getKey();
Action action = entry.getValue();
switch (action.type()) {
case DROP -> committer.prepareDropTable(schemaTableName);
case ALTER -> committer.prepareAlterTable(action.identity(), action.queryId(), action.data());
case ADD -> committer.prepareAddTable(action.identity(), action.queryId(), action.data());
case INSERT_EXISTING -> committer.prepareInsertExistingTable(action.identity(), action.queryId(), action.data());
case MERGE -> committer.prepareMergeExistingTable(action.identity(), action.data());
case DROP_PRESERVE_DATA -> throw new IllegalArgumentException("Unsupported action type: " + action.type());
}
}
for (Entry, Action>> tableEntry : partitionActions.entrySet()) {
SchemaTableName schemaTableName = tableEntry.getKey();
for (Entry, Action> partitionEntry : tableEntry.getValue().entrySet()) {
List partitionValues = partitionEntry.getKey();
Action action = partitionEntry.getValue();
switch (action.type()) {
case DROP -> committer.prepareDropPartition(schemaTableName, partitionValues, true);
case DROP_PRESERVE_DATA -> committer.prepareDropPartition(schemaTableName, partitionValues, false);
case ALTER -> committer.prepareAlterPartition(action.identity(), action.queryId(), action.data());
case ADD -> committer.prepareAddPartition(action.identity(), action.queryId(), action.data());
case INSERT_EXISTING, MERGE -> committer.prepareInsertExistingPartition(action.identity(), action.queryId(), action.data());
}
}
}
// Wait for all file system operations for "INSERT_EXISTING" and "ADD" action to finish
committer.waitForAsyncFileSystemOperations();
// At this point, all file system operations, whether asynchronously issued or not, have completed successfully.
// We are moving on to metastore operations now.
committer.executeAddTableOperations(transaction);
committer.executeAlterTableOperations();
committer.executeAlterPartitionOperations();
committer.executeAddPartitionOperations(transaction);
committer.executeUpdateStatisticsOperations(transaction);
}
catch (Throwable t) {
log.warn("Rolling back due to metastore commit failure: %s", t.getMessage());
try {
committer.cancelUnstartedAsyncFileSystemOperations();
committer.undoUpdateStatisticsOperations(transaction);
committer.undoAddPartitionOperations();
committer.undoAddTableOperations();
committer.waitForAsyncFileSystemOperationSuppressThrowable();
// fileSystemFutures must all come back before any file system cleanups are carried out.
// Otherwise, files that should be deleted may be created after cleanup is done.
committer.executeCleanupTasksForAbort(declaredIntentionsToWrite);
committer.executeRenameTasksForAbort();
// Partition directory must be put back before the relevant metastore operation can be undone
committer.undoAlterTableOperations();
committer.undoAlterPartitionOperations();
rollbackShared();
}
catch (RuntimeException e) {
t.addSuppressed(new Exception("Failed to roll back after commit failure", e));
}
throw t;
}
finally {
committer.executeTableInvalidationCallback();
}
try {
// After this line, operations are no longer reversible.
// The next section will deal with "dropping table/partition". Commit may still fail in
// this section. Even if the commit fails, cleanups, instead of rollbacks, will be executed.
committer.executeIrreversibleMetastoreOperations();
// If control flow reached this point, this commit is considered successful no matter
// what happens later. The only operations that haven't been carried out yet
// are cleanup operations.
// The program control flow will go to finally next. And cleanup will run because
// moveForwardInFinally has been set to false.
}
finally {
// In this method, all operations are best-effort cleanup operations.
// If any operation fails, the error will be logged and ignored.
// Additionally, other cleanup operations should still be attempted.
// Execute deletion tasks
committer.executeDeletionTasksForFinish();
// Clean up staging directories (that may recursively contain empty directories or stale files from failed attempts)
committer.pruneAndDeleteStagingDirectories(declaredIntentionsToWrite);
}
}
private class Committer
{
private final AtomicBoolean fileSystemOperationsCancelled = new AtomicBoolean(false);
private final List> fileSystemOperationFutures = new ArrayList<>();
// File system
// For file system changes, only operations outside the writing paths (as specified in declared intentions to write)
// need to MOVE_BACKWARD tasks scheduled. Files in writing paths are handled by rollbackShared().
private final List deletionTasksForFinish = new ArrayList<>();
private final List renameTasksForAbort = new ArrayList<>();
private final Queue cleanUpTasksForAbort = new ConcurrentLinkedQueue<>();
// Notify callback about changes on the schema tables / partitions
private final Set
tablesToInvalidate = new LinkedHashSet<>();
private final Set partitionsToInvalidate = new LinkedHashSet<>();
// Metastore
private final List addTableOperations = new ArrayList<>();
private final List alterTableOperations = new ArrayList<>();
private final Map partitionAdders = new HashMap<>();
private final List alterPartitionOperations = new ArrayList<>();
private final List updateStatisticsOperations = new ArrayList<>();
private final List metastoreDeleteOperations = new ArrayList<>();
private final AcidTransaction transaction;
// Flag for better error message
private boolean deleteOnly = true;
Committer(AcidTransaction transaction)
{
this.transaction = transaction;
}
private void prepareDropTable(SchemaTableName schemaTableName)
{
metastoreDeleteOperations.add(new IrreversibleMetastoreOperation(
format("drop table %s", schemaTableName),
() -> {
Optional
droppedTable = delegate.getTable(schemaTableName.getSchemaName(), schemaTableName.getTableName());
try {
delegate.dropTable(schemaTableName.getSchemaName(), schemaTableName.getTableName(), true);
}
finally {
// perform explicit invalidation for the table in irreversible metastore operation
droppedTable.ifPresent(tableInvalidationCallback::invalidate);
}
}));
}
private void prepareAlterTable(ConnectorIdentity identity, String queryId, TableAndMore tableAndMore)
{
deleteOnly = false;
Table table = tableAndMore.getTable();
Location targetLocation = Location.of(table.getStorage().getLocation());
Table oldTable = delegate.getTable(table.getDatabaseName(), table.getTableName())
.orElseThrow(() -> new TrinoException(TRANSACTION_CONFLICT, "The table that this transaction modified was deleted in another transaction. " + table.getSchemaTableName()));
Location oldTableLocation = Location.of(oldTable.getStorage().getLocation());
tablesToInvalidate.add(oldTable);
cleanExtraOutputFiles(identity, queryId, tableAndMore);
// The location of the old table and the new table can be different because we allow arbitrary directories through LocationService.
// If the location of the old table is the same as the location of the new table:
// * Rename the old data directory to a temporary path with a special suffix
// * Remember we will need to delete that directory at the end if transaction successfully commits
// * Remember we will need to undo the rename if transaction aborts
// Otherwise,
// * Remember we will need to delete the location of the old partition at the end if transaction successfully commits
if (targetLocation.equals(oldTableLocation)) {
Location location = asFileLocation(oldTableLocation);
Location oldTableStagingPath = location.parentDirectory().appendPath("_temp_" + location.fileName() + "_" + queryId);
renameDirectory(
fileSystemFactory.create(identity),
oldTableLocation,
oldTableStagingPath,
() -> renameTasksForAbort.add(new DirectoryRenameTask(identity, oldTableStagingPath, oldTableLocation)));
if (!skipDeletionForAlter) {
deletionTasksForFinish.add(new DirectoryDeletionTask(identity, oldTableStagingPath));
}
}
else {
if (!skipDeletionForAlter) {
deletionTasksForFinish.add(new DirectoryDeletionTask(identity, oldTableLocation));
}
}
Location currentLocation = tableAndMore.getCurrentLocation()
.orElseThrow(() -> new IllegalArgumentException("location should be present for alter table"));
if (!targetLocation.equals(currentLocation)) {
renameDirectory(
fileSystemFactory.create(identity),
currentLocation,
targetLocation,
() -> cleanUpTasksForAbort.add(new DirectoryCleanUpTask(identity, targetLocation, true)));
}
// Partition alter must happen regardless of whether the original and current location is the same
// because metadata might change: e.g., storage format, column types, etc.
alterTableOperations.add(new AlterTableOperation(tableAndMore.getTable(), oldTable, tableAndMore.getPrincipalPrivileges()));
updateStatisticsOperations.add(new UpdateStatisticsOperation(
table.getSchemaTableName(),
Optional.empty(),
tableAndMore.getStatisticsUpdate(),
false));
}
private void prepareAddTable(ConnectorIdentity identity, String queryId, TableAndMore tableAndMore)
{
deleteOnly = false;
cleanExtraOutputFiles(identity, queryId, tableAndMore);
Table table = tableAndMore.getTable();
if (table.getTableType().equals(MANAGED_TABLE.name())) {
Optional targetLocation = table.getStorage().getOptionalLocation().map(Location::of);
if (targetLocation.isPresent()) {
Optional currentLocation = tableAndMore.getCurrentLocation();
Location targetPath = targetLocation.get();
TrinoFileSystem fileSystem = fileSystemFactory.create(identity);
if (table.getPartitionColumns().isEmpty() && currentLocation.isPresent()) {
// CREATE TABLE AS SELECT unpartitioned table
if (targetPath.equals(currentLocation.get())) {
// Target path and current path are the same. Therefore, directory move is not needed.
}
else {
renameDirectory(
fileSystem,
currentLocation.get(),
targetPath,
() -> cleanUpTasksForAbort.add(new DirectoryCleanUpTask(identity, targetPath, true)));
}
}
else {
// CREATE TABLE AS SELECT partitioned table, or
// CREATE TABLE partitioned/unpartitioned table (without data)
if (directoryExists(fileSystem, targetPath)) {
if (currentLocation.isPresent() && currentLocation.get().equals(targetPath)) {
// It is okay to skip directory creation when currentLocation is equal to targetPath
// because the directory may have been created when creating partition directories.
// However, it is important to note that the two being equal does not guarantee
// a directory had been created.
}
else {
throw new TrinoException(
HIVE_PATH_ALREADY_EXISTS,
format("Unable to create directory %s: target directory already exists", targetPath));
}
}
else {
cleanUpTasksForAbort.add(new DirectoryCleanUpTask(identity, targetPath, true));
createDirectory(fileSystem, targetPath);
}
}
}
// if targetLocation is not set in table, we assume HMS creates table directory
}
addTableOperations.add(new CreateTableOperation(table, tableAndMore.getPrincipalPrivileges(), tableAndMore.isIgnoreExisting(), tableAndMore.getStatisticsUpdate()));
}
private void prepareInsertExistingTable(ConnectorIdentity identity, String queryId, TableAndMore tableAndMore)
{
deleteOnly = false;
Table table = tableAndMore.getTable();
Location targetPath = Location.of(table.getStorage().getLocation());
tablesToInvalidate.add(table);
Location currentPath = tableAndMore.getCurrentLocation().orElseThrow();
cleanUpTasksForAbort.add(new DirectoryCleanUpTask(identity, targetPath, false));
if (!targetPath.equals(currentPath)) {
// if staging directory is used, we cherry-pick files to be moved
TrinoFileSystem fileSystem = fileSystemFactory.create(identity);
asyncRename(fileSystem, fileSystemExecutor, fileSystemOperationsCancelled, fileSystemOperationFutures, currentPath, targetPath, tableAndMore.getFileNames().orElseThrow());
}
else {
// if we inserted directly into table directory, we need to remove extra output files which should not be part of the table
cleanExtraOutputFiles(identity, queryId, tableAndMore);
}
updateStatisticsOperations.add(new UpdateStatisticsOperation(
table.getSchemaTableName(),
Optional.empty(),
tableAndMore.getStatisticsUpdate(),
true));
if (isAcidTransactionRunning()) {
AcidTransaction transaction = getRequiredAcidTransaction();
updateTableWriteId(table.getDatabaseName(), table.getTableName(), transaction.getAcidTransactionId(), transaction.getWriteId(), OptionalLong.empty());
}
}
private void prepareMergeExistingTable(ConnectorIdentity identity, TableAndMore tableAndMore)
{
AcidTransaction transaction = getRequiredAcidTransaction();
checkArgument(transaction.isMerge(), "transaction should be merge, but is %s", transaction);
deleteOnly = false;
Table table = tableAndMore.getTable();
Location targetPath = Location.of(table.getStorage().getLocation());
Location currentPath = tableAndMore.getCurrentLocation().orElseThrow();
cleanUpTasksForAbort.add(new DirectoryCleanUpTask(identity, targetPath, false));
if (!targetPath.equals(currentPath)) {
TrinoFileSystem fileSystem = fileSystemFactory.create(identity);
asyncRename(fileSystem, fileSystemExecutor, fileSystemOperationsCancelled, fileSystemOperationFutures, currentPath, targetPath, tableAndMore.getFileNames().orElseThrow());
}
updateStatisticsOperations.add(new UpdateStatisticsOperation(
table.getSchemaTableName(),
Optional.empty(),
tableAndMore.getStatisticsUpdate(),
true));
updateTableWriteId(table.getDatabaseName(), table.getTableName(), transaction.getAcidTransactionId(), transaction.getWriteId(), OptionalLong.empty());
}
private void prepareDropPartition(SchemaTableName schemaTableName, List partitionValues, boolean deleteData)
{
metastoreDeleteOperations.add(new IrreversibleMetastoreOperation(
format("drop partition %s.%s %s", schemaTableName.getSchemaName(), schemaTableName.getTableName(), partitionValues),
() -> {
Optional droppedPartition = getOptionalPartition(delegate, schemaTableName, partitionValues);
try {
delegate.dropPartition(schemaTableName.getSchemaName(), schemaTableName.getTableName(), partitionValues, deleteData);
}
finally {
// perform explicit invalidation for the partition in irreversible metastore operation
droppedPartition.ifPresent(tableInvalidationCallback::invalidate);
}
}));
}
private void prepareAlterPartition(ConnectorIdentity identity, String queryId, PartitionAndMore partitionAndMore)
{
deleteOnly = false;
Partition partition = partitionAndMore.partition();
partitionsToInvalidate.add(partition);
String targetLocation = partition.getStorage().getLocation();
Partition oldPartition = getOptionalPartition(delegate, partition.getSchemaTableName(), partition.getValues())
.orElseThrow(() -> new TrinoException(
TRANSACTION_CONFLICT,
format("The partition that this transaction modified was deleted in another transaction. %s %s", partition.getTableName(), partition.getValues())));
String partitionName = getPartitionName(partition.getDatabaseName(), partition.getTableName(), partition.getValues());
PartitionStatistics oldPartitionStatistics = getExistingPartitionStatistics(partition, partitionName);
String oldPartitionLocation = oldPartition.getStorage().getLocation();
Location oldPartitionPath = asFileLocation(Location.of(oldPartitionLocation));
cleanExtraOutputFiles(identity, queryId, partitionAndMore);
// The location of the old partition and the new partition can be different because we allow arbitrary directories through LocationService.
// If the location of the old partition is the same as the location of the new partition:
// * Rename the old data directory to a temporary path with a special suffix
// * Remember we will need to delete that directory at the end if transaction successfully commits
// * Remember we will need to undo the rename if transaction aborts
// Otherwise,
// * Remember we will need to delete the location of the old partition at the end if transaction successfully commits
if (targetLocation.equals(oldPartitionLocation)) {
Location oldPartitionStagingPath = oldPartitionPath.sibling("_temp_" + oldPartitionPath.fileName() + "_" + queryId);
renameDirectory(
fileSystemFactory.create(identity),
oldPartitionPath,
oldPartitionStagingPath,
() -> renameTasksForAbort.add(new DirectoryRenameTask(identity, oldPartitionStagingPath, oldPartitionPath)));
if (!skipDeletionForAlter) {
deletionTasksForFinish.add(new DirectoryDeletionTask(identity, oldPartitionStagingPath));
}
}
else {
if (!skipDeletionForAlter) {
deletionTasksForFinish.add(new DirectoryDeletionTask(identity, oldPartitionPath));
}
}
Location currentPath = partitionAndMore.currentLocation();
Location targetPath = Location.of(targetLocation);
if (!targetPath.equals(currentPath)) {
renameDirectory(
fileSystemFactory.create(identity),
currentPath,
targetPath,
() -> cleanUpTasksForAbort.add(new DirectoryCleanUpTask(identity, targetPath, true)));
}
// Partition alter must happen regardless of whether the original and current location is the same
// because metadata might change: e.g., storage format, column types, etc.
alterPartitionOperations.add(new AlterPartitionOperation(
new PartitionWithStatistics(partition, partitionName, partitionAndMore.statisticsUpdate()),
new PartitionWithStatistics(oldPartition, partitionName, oldPartitionStatistics)));
}
private void cleanExtraOutputFiles(ConnectorIdentity identity, String queryId, PartitionAndMore partitionAndMore)
{
if (!partitionAndMore.cleanExtraOutputFilesOnCommit()) {
return;
}
verify(partitionAndMore.hasFileNames(), "fileNames expected to be set if isCleanExtraOutputFilesOnCommit is true");
TrinoFileSystem fileSystem = fileSystemFactory.create(identity);
SemiTransactionalHiveMetastore.cleanExtraOutputFiles(fileSystem, queryId, partitionAndMore.currentLocation(), ImmutableSet.copyOf(partitionAndMore.getFileNames()));
}
private void cleanExtraOutputFiles(ConnectorIdentity identity, String queryId, TableAndMore tableAndMore)
{
if (!tableAndMore.isCleanExtraOutputFilesOnCommit()) {
return;
}
TrinoFileSystem fileSystem = fileSystemFactory.create(identity);
Location tableLocation = tableAndMore.getCurrentLocation().orElseThrow(() ->
new IllegalArgumentException("currentLocation expected to be set if isCleanExtraOutputFilesOnCommit is true"));
List files = tableAndMore.getFileNames().orElseThrow(() -> new IllegalArgumentException("fileNames expected to be set if isCleanExtraOutputFilesOnCommit is true"));
SemiTransactionalHiveMetastore.cleanExtraOutputFiles(fileSystem, queryId, tableLocation, ImmutableSet.copyOf(files));
}
private PartitionStatistics getExistingPartitionStatistics(Partition partition, String partitionName)
{
try {
HiveBasicStatistics basicStatistics = getHiveBasicStatistics(partition.getParameters());
Map columnStatistics = delegate.getPartitionColumnStatistics(
partition.getDatabaseName(),
partition.getTableName(),
ImmutableSet.of(partitionName),
partition.getColumns().stream().map(Column::getName).collect(toImmutableSet()))
.get(partitionName);
if (columnStatistics == null) {
throw new TrinoException(
TRANSACTION_CONFLICT,
format("The partition that this transaction modified was deleted in another transaction. %s %s", partition.getTableName(), partition.getValues()));
}
return new PartitionStatistics(basicStatistics, columnStatistics);
}
catch (TrinoException e) {
if (e.getErrorCode().equals(HIVE_CORRUPTED_COLUMN_STATISTICS.toErrorCode())) {
log.warn(
e,
"Corrupted statistics found when altering partition. Table: %s.%s. Partition: %s",
partition.getDatabaseName(),
partition.getTableName(),
partition.getValues());
return PartitionStatistics.empty();
}
throw e;
}
}
private void prepareAddPartition(ConnectorIdentity identity, String queryId, PartitionAndMore partitionAndMore)
{
deleteOnly = false;
Partition partition = partitionAndMore.partition();
String targetLocation = partition.getStorage().getLocation();
Location currentPath = partitionAndMore.currentLocation();
Location targetPath = Location.of(targetLocation);
cleanExtraOutputFiles(identity, queryId, partitionAndMore);
PartitionAdder partitionAdder = partitionAdders.computeIfAbsent(
partition.getSchemaTableName(),
_ -> new PartitionAdder(partition.getDatabaseName(), partition.getTableName(), delegate, PARTITION_COMMIT_BATCH_SIZE));
fileSystemOperationFutures.add(CompletableFuture.runAsync(() -> {
if (fileSystemOperationsCancelled.get()) {
return;
}
TrinoFileSystem fileSystem = fileSystemFactory.create(identity);
if (directoryExists(fileSystem, currentPath)) {
if (!targetPath.equals(currentPath)) {
renameDirectory(
fileSystem,
currentPath,
targetPath,
() -> cleanUpTasksForAbort.add(new DirectoryCleanUpTask(identity, targetPath, true)));
}
}
else {
cleanUpTasksForAbort.add(new DirectoryCleanUpTask(identity, targetPath, true));
createDirectory(fileSystem, targetPath);
}
}, fileSystemExecutor));
String partitionName = getPartitionName(partition.getDatabaseName(), partition.getTableName(), partition.getValues());
partitionAdder.addPartition(new PartitionWithStatistics(partition, partitionName, partitionAndMore.statisticsUpdate()));
}
private void prepareInsertExistingPartition(ConnectorIdentity identity, String queryId, PartitionAndMore partitionAndMore)
{
deleteOnly = false;
Partition partition = partitionAndMore.partition();
partitionsToInvalidate.add(partition);
Location targetPath = Location.of(partition.getStorage().getLocation());
Location currentPath = partitionAndMore.currentLocation();
cleanUpTasksForAbort.add(new DirectoryCleanUpTask(identity, targetPath, false));
if (!targetPath.equals(currentPath)) {
// if staging directory is used, we cherry-pick files to be moved
TrinoFileSystem fileSystem = fileSystemFactory.create(identity);
asyncRename(fileSystem, fileSystemExecutor, fileSystemOperationsCancelled, fileSystemOperationFutures, currentPath, targetPath, partitionAndMore.getFileNames());
}
else {
// if we inserted directly into partition directory, we need to remove extra output files which should not be part of the table
cleanExtraOutputFiles(identity, queryId, partitionAndMore);
}
updateStatisticsOperations.add(new UpdateStatisticsOperation(
partition.getSchemaTableName(),
Optional.of(getPartitionName(partition.getDatabaseName(), partition.getTableName(), partition.getValues())),
partitionAndMore.statisticsUpdate(),
true));
}
private void executeCleanupTasksForAbort(Collection declaredIntentionsToWrite)
{
Set queryIds = declaredIntentionsToWrite.stream()
.map(DeclaredIntentionToWrite::queryId)
.collect(toImmutableSet());
for (DirectoryCleanUpTask cleanUpTask : cleanUpTasksForAbort) {
recursiveDeleteFilesAndLog(cleanUpTask.identity(), cleanUpTask.location(), queryIds, cleanUpTask.deleteEmptyDirectory(), "temporary directory commit abort");
}
}
private void executeDeletionTasksForFinish()
{
for (DirectoryDeletionTask deletionTask : deletionTasksForFinish) {
TrinoFileSystem fileSystem = fileSystemFactory.create(deletionTask.identity());
try {
fileSystem.deleteDirectory(deletionTask.location());
}
catch (IOException e) {
logCleanupFailure(e, "Error deleting directory: %s", deletionTask.location());
}
}
}
private void executeRenameTasksForAbort()
{
for (DirectoryRenameTask directoryRenameTask : renameTasksForAbort) {
try {
// Ignore the task if the source directory doesn't exist.
// This is probably because the original rename that we are trying to undo here never succeeded.
TrinoFileSystem fileSystem = fileSystemFactory.create(directoryRenameTask.identity());
if (directoryExists(fileSystem, directoryRenameTask.renameFrom())) {
renameDirectory(fileSystem, directoryRenameTask.renameFrom(), directoryRenameTask.renameTo(), () -> {});
}
}
catch (Throwable throwable) {
logCleanupFailure(throwable, "failed to undo rename of partition directory: %s to %s", directoryRenameTask.renameFrom(), directoryRenameTask.renameTo());
}
}
}
private void pruneAndDeleteStagingDirectories(List declaredIntentionsToWrite)
{
for (DeclaredIntentionToWrite declaredIntentionToWrite : declaredIntentionsToWrite) {
if (declaredIntentionToWrite.mode() != WriteMode.STAGE_AND_MOVE_TO_TARGET_DIRECTORY) {
continue;
}
Set queryIds = declaredIntentionsToWrite.stream()
.map(DeclaredIntentionToWrite::queryId)
.collect(toImmutableSet());
Location path = declaredIntentionToWrite.rootPath();
recursiveDeleteFilesAndLog(declaredIntentionToWrite.identity(), path, queryIds, true, "staging directory cleanup");
}
}
private void waitForAsyncFileSystemOperations()
{
for (CompletableFuture future : fileSystemOperationFutures) {
getFutureValue(future, TrinoException.class);
}
}
private void waitForAsyncFileSystemOperationSuppressThrowable()
{
for (CompletableFuture future : fileSystemOperationFutures) {
try {
future.get();
}
catch (InterruptedException e) {
Thread.currentThread().interrupt();
}
catch (Throwable t) {
// ignore
}
}
}
private void cancelUnstartedAsyncFileSystemOperations()
{
fileSystemOperationsCancelled.set(true);
}
private void executeAddTableOperations(AcidTransaction transaction)
{
for (CreateTableOperation addTableOperation : addTableOperations) {
addTableOperation.run(delegate, transaction);
}
}
private void executeAlterTableOperations()
{
for (AlterTableOperation alterTableOperation : alterTableOperations) {
alterTableOperation.run(delegate, transaction);
}
}
private void executeAlterPartitionOperations()
{
for (AlterPartitionOperation alterPartitionOperation : alterPartitionOperations) {
alterPartitionOperation.run(delegate);
}
}
private void executeAddPartitionOperations(AcidTransaction transaction)
{
for (PartitionAdder partitionAdder : partitionAdders.values()) {
partitionAdder.execute(transaction);
}
}
private void executeUpdateStatisticsOperations(AcidTransaction transaction)
{
ImmutableList.Builder> executeUpdateFutures = ImmutableList.builder();
List failedUpdateStatisticsOperationDescriptions = new ArrayList<>();
List suppressedExceptions = new ArrayList<>();
for (UpdateStatisticsOperation operation : updateStatisticsOperations) {
executeUpdateFutures.add(CompletableFuture.runAsync(() -> {
try {
operation.run(delegate, transaction);
}
catch (Throwable t) {
synchronized (failedUpdateStatisticsOperationDescriptions) {
addSuppressedExceptions(suppressedExceptions, t, failedUpdateStatisticsOperationDescriptions, operation.getDescription());
}
}
}, updateExecutor));
}
for (CompletableFuture executeUpdateFuture : executeUpdateFutures.build()) {
getFutureValue(executeUpdateFuture);
}
if (!suppressedExceptions.isEmpty()) {
StringBuilder message = new StringBuilder();
message.append("All operations other than the following update operations were completed: ");
Joiner.on("; ").appendTo(message, failedUpdateStatisticsOperationDescriptions);
TrinoException trinoException = new TrinoException(HIVE_METASTORE_ERROR, message.toString());
suppressedExceptions.forEach(trinoException::addSuppressed);
throw trinoException;
}
}
private void executeTableInvalidationCallback()
{
tablesToInvalidate.forEach(tableInvalidationCallback::invalidate);
partitionsToInvalidate.forEach(tableInvalidationCallback::invalidate);
}
private void undoAddPartitionOperations()
{
for (PartitionAdder partitionAdder : partitionAdders.values()) {
List> partitionsFailedToRollback = partitionAdder.rollback();
if (!partitionsFailedToRollback.isEmpty()) {
logCleanupFailure("Failed to rollback: add_partition for partitions %s.%s %s",
partitionAdder.getSchemaName(),
partitionAdder.getTableName(),
partitionsFailedToRollback);
}
}
}
private void undoAddTableOperations()
{
for (CreateTableOperation addTableOperation : addTableOperations) {
try {
addTableOperation.undo(delegate);
}
catch (Throwable throwable) {
logCleanupFailure(throwable, "failed to rollback: %s", addTableOperation.getDescription());
}
}
}
private void undoAlterTableOperations()
{
for (AlterTableOperation alterTableOperation : alterTableOperations) {
try {
alterTableOperation.undo(delegate, transaction);
}
catch (Throwable throwable) {
logCleanupFailure(throwable, "failed to rollback: %s", alterTableOperation.getDescription());
}
}
}
private void undoAlterPartitionOperations()
{
for (AlterPartitionOperation alterPartitionOperation : alterPartitionOperations) {
try {
alterPartitionOperation.undo(delegate);
}
catch (Throwable throwable) {
logCleanupFailure(throwable, "failed to rollback: %s", alterPartitionOperation.getDescription());
}
}
}
private void undoUpdateStatisticsOperations(AcidTransaction transaction)
{
ImmutableList.Builder> undoUpdateFutures = ImmutableList.builder();
for (UpdateStatisticsOperation operation : updateStatisticsOperations) {
undoUpdateFutures.add(CompletableFuture.runAsync(() -> {
try {
operation.undo(delegate, transaction);
}
catch (Throwable throwable) {
logCleanupFailure(throwable, "failed to rollback: %s", operation.getDescription());
}
}, updateExecutor));
}
for (CompletableFuture undoUpdateFuture : undoUpdateFutures.build()) {
getFutureValue(undoUpdateFuture);
}
}
private void executeIrreversibleMetastoreOperations()
{
List failedIrreversibleOperationDescriptions = new ArrayList<>();
List suppressedExceptions = new ArrayList<>();
AtomicBoolean anySucceeded = new AtomicBoolean(false);
ImmutableList.Builder> dropFutures = ImmutableList.builder();
for (IrreversibleMetastoreOperation irreversibleMetastoreOperation : metastoreDeleteOperations) {
dropFutures.add(CompletableFuture.runAsync(() -> {
try {
irreversibleMetastoreOperation.run();
anySucceeded.set(true);
}
catch (Throwable t) {
synchronized (failedIrreversibleOperationDescriptions) {
addSuppressedExceptions(suppressedExceptions, t, failedIrreversibleOperationDescriptions, irreversibleMetastoreOperation.description());
}
}
}, dropExecutor));
}
for (CompletableFuture dropFuture : dropFutures.build()) {
// none of the futures should fail because all exceptions are being handled explicitly
getFutureValue(dropFuture);
}
if (!suppressedExceptions.isEmpty()) {
StringBuilder message = new StringBuilder();
if (deleteOnly && !anySucceeded.get()) {
message.append("The following metastore delete operations failed: ");
}
else {
message.append("The transaction didn't commit cleanly. All operations other than the following delete operations were completed: ");
}
Joiner.on("; ").appendTo(message, failedIrreversibleOperationDescriptions);
TrinoException trinoException = new TrinoException(HIVE_METASTORE_ERROR, message.toString());
suppressedExceptions.forEach(trinoException::addSuppressed);
throw trinoException;
}
}
}
private static Optional getOptionalPartition(HiveMetastore metastore, SchemaTableName schemaTableName, List partitionValues)
{
return metastore.getTable(schemaTableName.getSchemaName(), schemaTableName.getTableName())
.flatMap(table -> metastore.getPartition(table, partitionValues));
}
@GuardedBy("this")
private void rollbackShared()
{
checkHoldsLock();
for (DeclaredIntentionToWrite declaredIntentionToWrite : declaredIntentionsToWrite) {
switch (declaredIntentionToWrite.mode()) {
case STAGE_AND_MOVE_TO_TARGET_DIRECTORY, DIRECT_TO_TARGET_NEW_DIRECTORY -> {
// For STAGE_AND_MOVE_TO_TARGET_DIRECTORY, there is no need to clean up the target directory as
// it will only be written to during the commit call and the commit call cleans up after failures.
if ((declaredIntentionToWrite.mode() == DIRECT_TO_TARGET_NEW_DIRECTORY) && skipTargetCleanupOnRollback) {
break;
}
Location rootPath = declaredIntentionToWrite.rootPath();
// In the case of DIRECT_TO_TARGET_NEW_DIRECTORY, if the directory is not guaranteed to be unique
// for the query, it is possible that another query or compute engine may see the directory, wrote
// data to it, and exported it through metastore. Therefore, it may be argued that cleanup of staging
// directories must be carried out conservatively. To be safe, we only delete files that start or
// end with the query IDs in this transaction.
recursiveDeleteFilesAndLog(
declaredIntentionToWrite.identity(),
rootPath,
ImmutableSet.of(declaredIntentionToWrite.queryId()),
true,
format("staging/target_new directory rollback for table %s", declaredIntentionToWrite.schemaTableName()));
}
case DIRECT_TO_TARGET_EXISTING_DIRECTORY -> {
Set pathsToClean = new HashSet<>();
// Check the base directory of the declared intention
// * existing partition may also be in this directory
// * this is where new partitions are created
Location baseDirectory = declaredIntentionToWrite.rootPath();
pathsToClean.add(baseDirectory);
SchemaTableName schemaTableName = declaredIntentionToWrite.schemaTableName();
Optional
table = delegate.getTable(schemaTableName.getSchemaName(), schemaTableName.getTableName());
if (table.isPresent()) {
// check every existing partition that is outside for the base directory
List partitionColumns = table.get().getPartitionColumns();
if (!partitionColumns.isEmpty()) {
List partitionColumnNames = partitionColumns.stream()
.map(Column::getName)
.collect(toImmutableList());
List partitionNames = getOptionalPartitions(
schemaTableName.getSchemaName(),
schemaTableName.getTableName(),
partitionColumnNames,
TupleDomain.all())
.orElse(ImmutableList.of());
for (List partitionNameBatch : Iterables.partition(partitionNames, 10)) {
Collection> partitions = getOptionalPartitions(schemaTableName.getSchemaName(), schemaTableName.getTableName(), partitionNameBatch).values();
partitions.stream()
.flatMap(Optional::stream)
.map(partition -> partition.getStorage().getLocation())
.filter(path -> !path.startsWith(baseDirectory.toString()))
.map(Location::of)
.forEach(pathsToClean::add);
}
}
}
else {
logCleanupFailure(
"Error rolling back write to table %s.%s. Data directory may contain temporary data. Table was dropped in another transaction.",
schemaTableName.getSchemaName(),
schemaTableName.getTableName());
}
// delete any file that starts or ends with the query ID
for (Location path : pathsToClean) {
// We cannot delete any of the directories here since we do not know who created them.
recursiveDeleteFilesAndLog(
declaredIntentionToWrite.identity(),
path,
ImmutableSet.of(declaredIntentionToWrite.queryId()),
false,
format("target_existing directory rollback for table %s", schemaTableName));
}
}
}
}
}
@VisibleForTesting
public synchronized void testOnlyCheckIsReadOnly()
{
if (state != State.EMPTY) {
throw new AssertionError("Test did not commit or rollback");
}
}
@GuardedBy("this")
private synchronized void checkReadable()
{
checkHoldsLock();
switch (state) {
case EMPTY, SHARED_OPERATION_BUFFERED -> {}
case EXCLUSIVE_OPERATION_BUFFERED -> throw new TrinoException(NOT_SUPPORTED, "Unsupported combination of operations in a single transaction");
case FINISHED -> throw new IllegalStateException("Tried to access metastore after transaction has been committed/aborted");
}
}
@GuardedBy("this")
private synchronized void setShared()
{
checkHoldsLock();
checkReadable();
state = State.SHARED_OPERATION_BUFFERED;
}
@GuardedBy("this")
private synchronized void setExclusive(ExclusiveOperation exclusiveOperation)
{
checkHoldsLock();
if (state != State.EMPTY) {
throw new TrinoException(NOT_SUPPORTED, "Unsupported combination of operations in a single transaction");
}
state = State.EXCLUSIVE_OPERATION_BUFFERED;
bufferedExclusiveOperation = exclusiveOperation;
}
@GuardedBy("this")
private void checkNoPartitionAction(String databaseName, String tableName)
{
checkHoldsLock();
Map, Action> partitionActionsOfTable = partitionActions.get(new SchemaTableName(databaseName, tableName));
if (partitionActionsOfTable != null && !partitionActionsOfTable.isEmpty()) {
throw new TrinoException(NOT_SUPPORTED, "Cannot make schema changes to a table/view with modified partitions in the same transaction");
}
}
@FormatMethod
private static void logCleanupFailure(String format, Object... args)
{
log.warn(format, args);
}
@FormatMethod
private static void logCleanupFailure(Throwable t, String format, Object... args)
{
log.warn(t, format, args);
}
private static void addSuppressedExceptions(List suppressedExceptions, Throwable t, List descriptions, String description)
{
descriptions.add(description);
// A limit is needed to avoid having a huge exception object. 5 was chosen arbitrarily.
if (suppressedExceptions.size() < 5) {
suppressedExceptions.add(t);
}
}
private static void asyncRename(
TrinoFileSystem fileSystem,
Executor executor,
AtomicBoolean cancelled,
List> fileRenameFutures,
Location currentPath,
Location targetPath,
List fileNames)
{
for (String fileName : fileNames) {
Location source = currentPath.appendPath(fileName);
Location target = targetPath.appendPath(fileName);
fileRenameFutures.add(CompletableFuture.runAsync(() -> {
if (cancelled.get()) {
return;
}
try {
fileSystem.renameFile(source, target);
}
catch (IOException e) {
throw new TrinoException(HIVE_FILESYSTEM_ERROR, format("Error moving data files from %s to final location %s", source, target), e);
}
}, executor));
}
}
private void recursiveDeleteFilesAndLog(ConnectorIdentity identity, Location directory, Set queryIds, boolean deleteEmptyDirectories, String reason)
{
RecursiveDeleteResult recursiveDeleteResult = recursiveDeleteFiles(
fileSystemFactory.create(identity),
directory,
queryIds,
deleteEmptyDirectories);
if (!recursiveDeleteResult.notDeletedEligibleItems().isEmpty()) {
logCleanupFailure(
"Error deleting directory %s for %s. Some eligible items cannot be deleted: %s.",
directory.toString(),
reason,
recursiveDeleteResult.notDeletedEligibleItems());
}
else if (deleteEmptyDirectories && !recursiveDeleteResult.directoryNoLongerExists()) {
logCleanupFailure(
"Error deleting directory %s for %s. Cannot delete the directory.",
directory.toString(),
reason);
}
}
/**
* Attempt to recursively remove eligible files and/or directories in {@code directory}.
*
* When {@code queryIds} is not present, all files (but not necessarily directories) will be
* ineligible. If all files shall be deleted, you can use an empty string as {@code queryIds}.
*
* When {@code deleteEmptySubDirectory} is true, any empty directory (including directories that
* were originally empty, and directories that become empty after files prefixed or suffixed with
* {@code queryIds} are deleted) will be eligible.
*
* This method will not delete anything that's neither a directory nor a file.
*
* @param queryIds prefix or suffix of files that should be deleted
* @param deleteEmptyDirectories whether empty directories should be deleted
*/
private static RecursiveDeleteResult recursiveDeleteFiles(TrinoFileSystem fileSystem, Location directory, Set queryIds, boolean deleteEmptyDirectories)
{
try {
if (!fileSystem.directoryExists(directory).orElse(false)) {
return new RecursiveDeleteResult(true, ImmutableList.of());
}
}
catch (IOException e) {
ImmutableList.Builder notDeletedItems = ImmutableList.builder();
notDeletedItems.add(directory.toString() + "/**");
return new RecursiveDeleteResult(false, notDeletedItems.build());
}
return doRecursiveDeleteFiles(fileSystem, directory, queryIds, deleteEmptyDirectories);
}
private static RecursiveDeleteResult doRecursiveDeleteFiles(TrinoFileSystem fileSystem, Location directory, Set queryIds, boolean deleteEmptyDirectories)
{
// don't delete hidden Trino directories use by FileHiveMetastore
directory = asFileLocation(directory);
if (directory.fileName().startsWith(".trino")) {
return new RecursiveDeleteResult(false, ImmutableList.of());
}
// TODO: this lists recursively but only uses the first level
List allFiles = new ArrayList<>();
Set allDirectories;
try {
FileIterator iterator = fileSystem.listFiles(directory);
while (iterator.hasNext()) {
Location location = iterator.next().location();
String child = location.toString().substring(directory.toString().length());
while (child.startsWith("/")) {
child = child.substring(1);
}
if (!child.contains("/")) {
allFiles.add(location);
}
}
allDirectories = fileSystem.listDirectories(directory);
}
catch (IOException e) {
ImmutableList.Builder notDeletedItems = ImmutableList.builder();
notDeletedItems.add(directory + "/**");
return new RecursiveDeleteResult(false, notDeletedItems.build());
}
boolean allDescendentsDeleted = true;
ImmutableList.Builder notDeletedEligibleItems = ImmutableList.builder();
for (Location file : allFiles) {
String fileName = file.fileName();
boolean eligible = false;
// don't delete hidden Trino directories use by FileHiveMetastore
if (!fileName.startsWith(".trino")) {
eligible = queryIds.stream().anyMatch(id -> isFileCreatedByQuery(fileName, id));
}
if (eligible) {
if (!deleteFileIfExists(fileSystem, file)) {
allDescendentsDeleted = false;
notDeletedEligibleItems.add(file.toString());
}
}
else {
allDescendentsDeleted = false;
}
}
for (Location file : allDirectories) {
RecursiveDeleteResult subResult = doRecursiveDeleteFiles(fileSystem, file, queryIds, deleteEmptyDirectories);
if (!subResult.directoryNoLongerExists()) {
allDescendentsDeleted = false;
}
if (!subResult.notDeletedEligibleItems().isEmpty()) {
notDeletedEligibleItems.addAll(subResult.notDeletedEligibleItems());
}
}
// Unconditionally delete empty delta_ and delete_delta_ directories, because that's
// what Hive does, and leaving them in place confuses delta file readers.
if (allDescendentsDeleted && (deleteEmptyDirectories || isDeltaDirectory(directory))) {
verify(notDeletedEligibleItems.build().isEmpty());
if (!deleteEmptyDirectoryIfExists(fileSystem, directory)) {
return new RecursiveDeleteResult(false, ImmutableList.of(directory + "/"));
}
return new RecursiveDeleteResult(true, ImmutableList.of());
}
return new RecursiveDeleteResult(false, notDeletedEligibleItems.build());
}
private static boolean isDeltaDirectory(Location directory)
{
return DELTA_DIRECTORY_MATCHER.matcher(asFileLocation(directory).fileName()).matches();
}
private static boolean deleteFileIfExists(TrinoFileSystem fileSystem, Location location)
{
try {
fileSystem.deleteFile(location);
return true;
}
catch (FileNotFoundException e) {
return true;
}
catch (IOException e) {
return false;
}
}
private static boolean deleteEmptyDirectoryIfExists(TrinoFileSystem fileSystem, Location location)
{
try {
if (fileSystem.listFiles(location).hasNext()) {
log.warn("Not deleting non-empty directory: %s", location);
return false;
}
fileSystem.deleteDirectory(location);
return true;
}
catch (IOException e) {
try {
return !fileSystem.directoryExists(location).orElse(false);
}
catch (IOException ex) {
return false;
}
}
}
private static void renameDirectory(TrinoFileSystem fileSystem, Location source, Location target, Runnable runWhenPathDoesntExist)
{
if (directoryExists(fileSystem, target)) {
throw new TrinoException(HIVE_PATH_ALREADY_EXISTS, format("Unable to rename from %s to %s: target directory already exists", source, target));
}
Location parent = asFileLocation(target).parentDirectory();
if (!directoryExists(fileSystem, parent)) {
createDirectory(fileSystem, parent);
}
// The runnable will assume that if rename fails, it will be okay to delete the directory (if the directory is empty).
// This is not technically true because a race condition still exists.
runWhenPathDoesntExist.run();
try {
fileSystem.renameDirectory(source, target);
}
catch (IOException e) {
throw new TrinoException(HIVE_FILESYSTEM_ERROR, format("Failed to rename %s to %s", source, target), e);
}
}
private static void createDirectory(TrinoFileSystem fileSystem, Location directory)
{
try {
fileSystem.createDirectory(directory);
}
catch (IOException e) {
throw new TrinoException(HIVE_FILESYSTEM_ERROR, e);
}
}
private static boolean directoryExists(TrinoFileSystem fileSystem, Location directory)
{
try {
return fileSystem.directoryExists(directory).orElse(false);
}
catch (IOException e) {
throw new TrinoException(HIVE_FILESYSTEM_ERROR, e);
}
}
private static Optional getQueryId(Database database)
{
return Optional.ofNullable(database.getParameters().get(TRINO_QUERY_ID_NAME));
}
private static Optional getQueryId(Table table)
{
return Optional.ofNullable(table.getParameters().get(TRINO_QUERY_ID_NAME));
}
private static Optional getQueryId(Partition partition)
{
return Optional.ofNullable(partition.getParameters().get(TRINO_QUERY_ID_NAME));
}
private static Location asFileLocation(Location location)
{
// TODO: this is to work around the file-only restriction of Location methods
String value = location.toString();
while (value.endsWith("/")) {
value = value.substring(0, value.length() - 1);
}
return Location.of(value);
}
private void checkHoldsLock()
{
// This method serves a similar purpose at runtime as GuardedBy on method serves during static analysis.
// This method should not have a significant performance impact. If it does, it may be reasonably to remove this method.
// This intentionally does not use checkState.
if (!Thread.holdsLock(this)) {
throw new IllegalStateException(format("Thread must hold a lock on the %s", getClass().getSimpleName()));
}
}
private enum State
{
EMPTY,
SHARED_OPERATION_BUFFERED,
EXCLUSIVE_OPERATION_BUFFERED,
FINISHED,
}
private enum ActionType
{
DROP,
DROP_PRESERVE_DATA,
ADD,
ALTER,
INSERT_EXISTING,
MERGE,
}
private enum TableSource
{
CREATED_IN_THIS_TRANSACTION,
PRE_EXISTING_TABLE,
// RECREATED_IN_THIS_TRANSACTION is a possible case, but it is not supported with the current implementation
}
private record Action(ActionType type, T data, ConnectorIdentity identity, String queryId)
{
private Action
{
requireNonNull(type, "type is null");
if (type == ActionType.DROP || type == ActionType.DROP_PRESERVE_DATA) {
checkArgument(data == null, "data is not null");
}
else {
requireNonNull(data, "data is null");
}
requireNonNull(identity, "identity is null");
requireNonNull(queryId, "queryId is null");
}
@Override
public T data()
{
checkState(type != ActionType.DROP);
return data;
}
}
private static class TableAndMore
{
private final Table table;
private final Optional principalPrivileges;
private final Optional currentLocation; // unpartitioned table only
private final Optional> fileNames;
private final boolean ignoreExisting;
private final PartitionStatistics statistics;
private final PartitionStatistics statisticsUpdate;
private final boolean cleanExtraOutputFilesOnCommit;
public TableAndMore(
Table table,
Optional principalPrivileges,
Optional currentLocation,
Optional> fileNames,
boolean ignoreExisting,
PartitionStatistics statistics,
PartitionStatistics statisticsUpdate,
boolean cleanExtraOutputFilesOnCommit)
{
this.table = requireNonNull(table, "table is null");
this.principalPrivileges = requireNonNull(principalPrivileges, "principalPrivileges is null");
this.currentLocation = requireNonNull(currentLocation, "currentLocation is null");
this.fileNames = requireNonNull(fileNames, "fileNames is null");
this.ignoreExisting = ignoreExisting;
this.statistics = requireNonNull(statistics, "statistics is null");
this.statisticsUpdate = requireNonNull(statisticsUpdate, "statisticsUpdate is null");
this.cleanExtraOutputFilesOnCommit = cleanExtraOutputFilesOnCommit;
checkArgument(!table.getStorage().getOptionalLocation().orElse("").isEmpty() || currentLocation.isEmpty(), "currentLocation cannot be supplied for table without location");
checkArgument(fileNames.isEmpty() || currentLocation.isPresent(), "fileNames can be supplied only when currentLocation is supplied");
}
public boolean isIgnoreExisting()
{
return ignoreExisting;
}
public Table getTable()
{
return table;
}
public PrincipalPrivileges getPrincipalPrivileges()
{
checkState(principalPrivileges.isPresent());
return principalPrivileges.get();
}
public Optional getCurrentLocation()
{
return currentLocation;
}
public Optional> getFileNames()
{
return fileNames;
}
public PartitionStatistics getStatistics()
{
return statistics;
}
public PartitionStatistics getStatisticsUpdate()
{
return statisticsUpdate;
}
public boolean isCleanExtraOutputFilesOnCommit()
{
return cleanExtraOutputFilesOnCommit;
}
@Override
public String toString()
{
return toStringHelper(this)
.add("table", table)
.add("principalPrivileges", principalPrivileges)
.add("currentLocation", currentLocation)
.add("fileNames", fileNames)
.add("ignoreExisting", ignoreExisting)
.add("statistics", statistics)
.add("statisticsUpdate", statisticsUpdate)
.add("cleanExtraOutputFilesOnCommit", cleanExtraOutputFilesOnCommit)
.toString();
}
}
private static class TableAndMergeResults
extends TableAndMore
{
private final List partitionMergeResults;
public TableAndMergeResults(Table table, Optional principalPrivileges, Optional currentLocation, List partitionMergeResults)
{
super(table, principalPrivileges, currentLocation, Optional.empty(), false, PartitionStatistics.empty(), PartitionStatistics.empty(), false); // retries are not supported for transactional tables
this.partitionMergeResults = requireNonNull(partitionMergeResults, "partitionMergeResults is null");
}
@Override
public String toString()
{
return toStringHelper(this)
.add("table", getTable())
.add("partitionMergeResults", partitionMergeResults)
.add("principalPrivileges", getPrincipalPrivileges())
.add("currentLocation", getCurrentLocation())
.toString();
}
}
private record PartitionAndMore(
Partition partition,
Location currentLocation,
Optional> fileNames,
PartitionStatistics statistics,
PartitionStatistics statisticsUpdate,
boolean cleanExtraOutputFilesOnCommit)
{
private PartitionAndMore
{
requireNonNull(partition, "partition is null");
requireNonNull(currentLocation, "currentLocation is null");
requireNonNull(fileNames, "fileNames is null");
requireNonNull(statistics, "statistics is null");
requireNonNull(statisticsUpdate, "statisticsUpdate is null");
}
public List getFileNames()
{
checkState(fileNames.isPresent());
return fileNames.get();
}
public boolean hasFileNames()
{
return fileNames.isPresent();
}
public Partition getAugmentedPartitionForInTransactionRead()
{
// This method augments the location field of the partition to the staging location.
// This way, if the partition is accessed in an ongoing transaction, staged data
// can be found and accessed.
Partition partition = this.partition;
String currentLocation = this.currentLocation.toString();
if (!currentLocation.equals(partition.getStorage().getLocation())) {
partition = Partition.builder(partition)
.withStorage(storage -> storage.setLocation(currentLocation))
.build();
}
return partition;
}
}
public record DeclaredIntentionToWrite(String declarationId, WriteMode mode, ConnectorIdentity identity, String queryId, Location rootPath, SchemaTableName schemaTableName)
{
public DeclaredIntentionToWrite
{
requireNonNull(declarationId, "declarationId is null");
requireNonNull(mode, "mode is null");
requireNonNull(identity, "identity is null");
requireNonNull(queryId, "queryId is null");
requireNonNull(rootPath, "rootPath is null");
requireNonNull(schemaTableName, "schemaTableName is null");
}
}
private record DirectoryCleanUpTask(ConnectorIdentity identity, Location location, boolean deleteEmptyDirectory)
{
private DirectoryCleanUpTask
{
requireNonNull(identity, "identity is null");
requireNonNull(location, "location is null");
}
}
private record DirectoryDeletionTask(ConnectorIdentity identity, Location location)
{
private DirectoryDeletionTask
{
requireNonNull(identity, "identity is null");
requireNonNull(location, "location is null");
}
}
private record DirectoryRenameTask(ConnectorIdentity identity, Location renameFrom, Location renameTo)
{
private DirectoryRenameTask
{
requireNonNull(identity, "identity is null");
requireNonNull(renameFrom, "renameFrom is null");
requireNonNull(renameTo, "renameTo is null");
}
}
private record IrreversibleMetastoreOperation(String description, Runnable action)
{
private IrreversibleMetastoreOperation
{
requireNonNull(description, "description is null");
requireNonNull(action, "action is null");
}
public void run()
{
action.run();
}
}
private static class CreateTableOperation
{
private final Table newTable;
private final PrincipalPrivileges privileges;
private boolean tableCreated;
private final boolean ignoreExisting;
private final PartitionStatistics statistics;
private final String queryId;
public CreateTableOperation(Table newTable, PrincipalPrivileges privileges, boolean ignoreExisting, PartitionStatistics statistics)
{
requireNonNull(newTable, "newTable is null");
this.newTable = newTable;
this.privileges = requireNonNull(privileges, "privileges is null");
this.ignoreExisting = ignoreExisting;
this.statistics = requireNonNull(statistics, "statistics is null");
this.queryId = getQueryId(newTable).orElseThrow(() -> new IllegalArgumentException("Query id is not present"));
}
public String getDescription()
{
return format("add table %s.%s", newTable.getDatabaseName(), newTable.getTableName());
}
public void run(HiveMetastore metastore, AcidTransaction transaction)
{
boolean created = false;
try {
metastore.createTable(newTable, privileges);
created = true;
}
catch (RuntimeException e) {
RuntimeException failure = e;
try {
Optional
existingTable = metastore.getTable(newTable.getDatabaseName(), newTable.getTableName());
if (existingTable.isPresent()) {
Table table = existingTable.get();
Optional existingTableQueryId = getQueryId(table);
if (existingTableQueryId.isPresent() && existingTableQueryId.get().equals(queryId)) {
// ignore table if it was already created by the same query during retries
failure = null;
created = true;
}
else {
// If the table definition in the metastore is different from what this tx wants to create,
// then there is a conflict (e.g., current tx wants to create T(a: bigint),
// but another tx already created T(a: varchar)).
// This may be a problem if there is an insert after this step.
if (!hasTheSameSchema(newTable, table)) {
// produce an understandable error message
failure = new TrinoException(TRANSACTION_CONFLICT, format("Table already exists with a different schema: '%s'", newTable.getTableName()));
}
else if (ignoreExisting) {
// if the statement is "CREATE TABLE IF NOT EXISTS", then ignore the exception
failure = null;
}
}
}
}
catch (RuntimeException _) {
// When table could not be fetched from metastore, it is not known whether the table was added.
// Deleting the table when aborting commit has the risk of deleting a table not added in this transaction.
// Not deleting the table may leave garbage behind. The former is much more dangerous than the latter.
// Therefore, the table is not considered added.
}
if (failure != null) {
throw failure;
}
}
tableCreated = true;
if (created && !isTrinoView(newTable) && !isTrinoMaterializedView(newTable)) {
metastore.updateTableStatistics(newTable.getDatabaseName(), newTable.getTableName(), transaction.getOptionalWriteId(), OVERWRITE_ALL, statistics);
}
}
private static boolean hasTheSameSchema(Table newTable, Table existingTable)
{
List newTableColumns = newTable.getDataColumns();
List existingTableColumns = existingTable.getDataColumns();
if (newTableColumns.size() != existingTableColumns.size()) {
return false;
}
for (Column existingColumn : existingTableColumns) {
if (newTableColumns.stream()
.noneMatch(newColumn -> newColumn.getName().equals(existingColumn.getName())
&& newColumn.getType().equals(existingColumn.getType()))) {
return false;
}
}
return true;
}
public void undo(HiveMetastore metastore)
{
if (!tableCreated) {
return;
}
metastore.dropTable(newTable.getDatabaseName(), newTable.getTableName(), false);
}
}
private static class AlterTableOperation
{
private final Table newTable;
private final Table oldTable;
private final PrincipalPrivileges principalPrivileges;
private boolean undo;
public AlterTableOperation(Table newTable, Table oldTable, PrincipalPrivileges principalPrivileges)
{
this.newTable = requireNonNull(newTable, "newTable is null");
this.oldTable = requireNonNull(oldTable, "oldTable is null");
this.principalPrivileges = requireNonNull(principalPrivileges, "principalPrivileges is null");
checkArgument(newTable.getDatabaseName().equals(oldTable.getDatabaseName()));
checkArgument(newTable.getTableName().equals(oldTable.getTableName()));
}
public String getDescription()
{
return format(
"alter table %s.%s",
newTable.getDatabaseName(),
newTable.getTableName());
}
public void run(HiveMetastore metastore, AcidTransaction transaction)
{
undo = true;
if (transaction.isTransactional()) {
metastore.alterTransactionalTable(newTable, transaction.getAcidTransactionId(), transaction.getWriteId(), principalPrivileges);
}
else {
metastore.replaceTable(newTable.getDatabaseName(), newTable.getTableName(), newTable, principalPrivileges);
}
}
public void undo(HiveMetastore metastore, AcidTransaction transaction)
{
if (!undo) {
return;
}
if (transaction.isTransactional()) {
metastore.alterTransactionalTable(oldTable, transaction.getAcidTransactionId(), transaction.getWriteId(), principalPrivileges);
}
else {
metastore.replaceTable(oldTable.getDatabaseName(), oldTable.getTableName(), oldTable, principalPrivileges);
}
}
}
private static class AlterPartitionOperation
{
private final PartitionWithStatistics newPartition;
private final PartitionWithStatistics oldPartition;
private boolean undo;
public AlterPartitionOperation(PartitionWithStatistics newPartition, PartitionWithStatistics oldPartition)
{
this.newPartition = requireNonNull(newPartition, "newPartition is null");
this.oldPartition = requireNonNull(oldPartition, "oldPartition is null");
checkArgument(newPartition.getPartition().getDatabaseName().equals(oldPartition.getPartition().getDatabaseName()));
checkArgument(newPartition.getPartition().getTableName().equals(oldPartition.getPartition().getTableName()));
checkArgument(newPartition.getPartition().getValues().equals(oldPartition.getPartition().getValues()));
}
public String getDescription()
{
return format(
"alter partition %s.%s %s",
newPartition.getPartition().getDatabaseName(),
newPartition.getPartition().getTableName(),
newPartition.getPartition().getValues());
}
public void run(HiveMetastore metastore)
{
undo = true;
metastore.alterPartition(newPartition.getPartition().getDatabaseName(), newPartition.getPartition().getTableName(), newPartition);
}
public void undo(HiveMetastore metastore)
{
if (!undo) {
return;
}
metastore.alterPartition(oldPartition.getPartition().getDatabaseName(), oldPartition.getPartition().getTableName(), oldPartition);
}
}
private static class UpdateStatisticsOperation
{
private final SchemaTableName tableName;
private final Optional partitionName;
private final PartitionStatistics statistics;
private final boolean merge;
private boolean done;
public UpdateStatisticsOperation(SchemaTableName tableName, Optional partitionName, PartitionStatistics statistics, boolean merge)
{
this.tableName = requireNonNull(tableName, "tableName is null");
this.partitionName = requireNonNull(partitionName, "partitionName is null");
this.statistics = requireNonNull(statistics, "statistics is null");
this.merge = merge;
}
public void run(HiveMetastore metastore, AcidTransaction transaction)
{
StatisticsUpdateMode mode = merge ? MERGE_INCREMENTAL : OVERWRITE_ALL;
if (partitionName.isPresent()) {
metastore.updatePartitionStatistics(
metastore.getTable(tableName.getSchemaName(), tableName.getTableName())
.orElseThrow(() -> new TableNotFoundException(tableName)),
mode,
ImmutableMap.of(partitionName.get(), statistics));
}
else {
metastore.updateTableStatistics(tableName.getSchemaName(), tableName.getTableName(), transaction.getOptionalWriteId(), mode, statistics);
}
done = true;
}
public void undo(HiveMetastore metastore, AcidTransaction transaction)
{
if (!done) {
return;
}
if (partitionName.isPresent()) {
metastore.updatePartitionStatistics(
metastore.getTable(tableName.getSchemaName(), tableName.getTableName())
.orElseThrow(() -> new TableNotFoundException(tableName)),
UNDO_MERGE_INCREMENTAL,
ImmutableMap.of(partitionName.get(), statistics));
}
else {
metastore.updateTableStatistics(tableName.getSchemaName(), tableName.getTableName(), transaction.getOptionalWriteId(), UNDO_MERGE_INCREMENTAL, statistics);
}
}
public String getDescription()
{
if (partitionName.isPresent()) {
return format("replace partition parameters %s %s", tableName, partitionName.get());
}
return format("replace table parameters %s", tableName);
}
}
private static class PartitionAdder
{
private final String schemaName;
private final String tableName;
private final HiveMetastore metastore;
private final int batchSize;
private final List partitions;
private List> createdPartitionValues = new ArrayList<>();
public PartitionAdder(String schemaName, String tableName, HiveMetastore metastore, int batchSize)
{
this.schemaName = schemaName;
this.tableName = tableName;
this.metastore = metastore;
this.batchSize = batchSize;
this.partitions = new ArrayList<>(batchSize);
}
public String getSchemaName()
{
return schemaName;
}
public String getTableName()
{
return tableName;
}
public void addPartition(PartitionWithStatistics partition)
{
checkArgument(getQueryId(partition.getPartition()).isPresent());
partitions.add(partition);
}
public void execute(AcidTransaction transaction)
{
List> batchedPartitions = Lists.partition(partitions, batchSize);
for (List batch : batchedPartitions) {
try {
metastore.addPartitions(schemaName, tableName, batch);
for (PartitionWithStatistics partition : batch) {
createdPartitionValues.add(partition.getPartition().getValues());
}
}
catch (Throwable t) {
// Add partition to the created list conservatively.
// Some metastore implementations are known to violate the "all or none" guarantee for add_partitions call.
boolean batchCompletelyAdded = true;
for (PartitionWithStatistics partition : batch) {
try {
Optional remotePartition = getOptionalPartition(metastore, new SchemaTableName(schemaName, tableName), partition.getPartition().getValues());
// getQueryId(partition) is guaranteed to be non-empty. It is asserted in PartitionAdder.addPartition.
if (remotePartition.isPresent() && getQueryId(remotePartition.get()).equals(getQueryId(partition.getPartition()))) {
createdPartitionValues.add(partition.getPartition().getValues());
}
else {
batchCompletelyAdded = false;
}
}
catch (Throwable _) {
// When partition could not be fetched from metastore, it is not known whether the partition was added.
// Deleting the partition when aborting commit has the risk of deleting partition not added in this transaction.
// Not deleting the partition may leave garbage behind. The former is much more dangerous than the latter.
// Therefore, the partition is not added to the createdPartitionValues list here.
batchCompletelyAdded = false;
}
}
// If all the partitions were added successfully, the add_partition operation was actually successful.
// For some reason, it threw an exception (communication failure, retry failure after communication failure, etc.).
// But we would consider it successful regardless.
if (!batchCompletelyAdded) {
if (t instanceof TableNotFoundException) {
throw new TrinoException(HIVE_TABLE_DROPPED_DURING_QUERY, t);
}
throw t;
}
}
}
if (transaction.isAcidTransactionRunning()) {
List partitionNames = partitions.stream().map(PartitionWithStatistics::getPartitionName).toList();
metastore.addDynamicPartitions(schemaName, tableName, partitionNames, transaction.getAcidTransactionId(), transaction.getWriteId(), transaction.getOperation());
}
partitions.clear();
}
public List> rollback()
{
// drop created partitions
List> partitionsFailedToRollback = new ArrayList<>();
for (List createdPartitionValue : createdPartitionValues) {
try {
metastore.dropPartition(schemaName, tableName, createdPartitionValue, false);
}
catch (PartitionNotFoundException e) {
// Maybe someone deleted the partition we added.
// Anyway, we are good because the partition is not there anymore.
}
catch (Throwable t) {
partitionsFailedToRollback.add(createdPartitionValue);
}
}
createdPartitionValues = partitionsFailedToRollback;
return partitionsFailedToRollback;
}
}
private record RecursiveDeleteResult(boolean directoryNoLongerExists, List notDeletedEligibleItems) {}
private interface ExclusiveOperation
{
void execute(HiveMetastore delegate);
}
private long allocateWriteId(String dbName, String tableName, long transactionId)
{
return delegate.allocateWriteId(dbName, tableName, transactionId);
}
private void acquireTableWriteLock(
AcidTransactionOwner transactionOwner,
String queryId,
long transactionId,
String dbName,
String tableName,
AcidOperation operation,
boolean isPartitioned)
{
delegate.acquireTableWriteLock(transactionOwner, queryId, transactionId, dbName, tableName, operation, isPartitioned);
}
private void updateTableWriteId(String dbName, String tableName, long transactionId, long writeId, OptionalLong rowCountChange)
{
delegate.updateTableWriteId(dbName, tableName, transactionId, writeId, rowCountChange);
}
public void addDynamicPartitions(String dbName, String tableName, List partitionNames, long transactionId, long writeId, AcidOperation operation)
{
delegate.addDynamicPartitions(dbName, tableName, partitionNames, transactionId, writeId, operation);
}
public static void cleanExtraOutputFiles(TrinoFileSystem fileSystem, String queryId, Location path, Set filesToKeep)
{
List filesToDelete = new ArrayList<>();
try {
Failsafe.with(DELETE_RETRY_POLICY).run(() -> {
log.debug("Deleting failed attempt files from %s for query %s", path, queryId);
filesToDelete.clear();
FileIterator iterator = fileSystem.listFiles(path);
while (iterator.hasNext()) {
Location file = iterator.next().location();
if (isFileCreatedByQuery(file.fileName(), queryId) && !filesToKeep.contains(file.fileName())) {
filesToDelete.add(file);
}
}
log.debug("Found %s failed attempt file(s) to delete for query %s", filesToDelete.size(), queryId);
fileSystem.deleteFiles(filesToDelete);
});
}
catch (FailsafeException e) {
// If we fail here, the query will be rolled back. The optimal outcome would be for rollback to complete successfully and clean up everything for the query.
// Yet if we have a problem here, probably rollback will also fail.
//
// Thrown exception is listing files which we could not delete. So the user can clean up those later manually.
// Note it is not a bullet-proof solution.
// The rollback routine will still fire and try to clean up any changes the query made. It will clean up some, and probably leave some behind.
//
// Still, we cannot do much better for non-transactional Hive tables.
throw new TrinoException(
HIVE_FILESYSTEM_ERROR,
format("Error deleting failed retry attempt files from %s; remaining files %s; manual cleanup may be required", path, filesToDelete),
e);
}
}
public record PartitionUpdateInfo(List
