org.apache.phoenix.schema.Sequence Maven / Gradle / Ivy
The newest version!
/*
* Licensed to the Apache Software Foundation (ASF) under one
* or more contributor license agreements. See the NOTICE file
* distributed with this work for additional information
* regarding copyright ownership. The ASF licenses this file
* to you under the Apache License, Version 2.0 (the
* "License"); you may not use this file except in compliance
* with the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package org.apache.phoenix.schema;
import static org.apache.phoenix.jdbc.PhoenixDatabaseMetaData.CACHE_SIZE_BYTES;
import static org.apache.phoenix.jdbc.PhoenixDatabaseMetaData.CURRENT_VALUE_BYTES;
import static org.apache.phoenix.jdbc.PhoenixDatabaseMetaData.CYCLE_FLAG_BYTES;
import static org.apache.phoenix.jdbc.PhoenixDatabaseMetaData.INCREMENT_BY_BYTES;
import static org.apache.phoenix.jdbc.PhoenixDatabaseMetaData.LIMIT_REACHED_FLAG_BYTES;
import static org.apache.phoenix.jdbc.PhoenixDatabaseMetaData.MAX_VALUE_BYTES;
import static org.apache.phoenix.jdbc.PhoenixDatabaseMetaData.MIN_VALUE_BYTES;
import static org.apache.phoenix.jdbc.PhoenixDatabaseMetaData.SYSTEM_SEQUENCE_FAMILY_BYTES;
import java.io.IOException;
import java.sql.SQLException;
import java.util.Arrays;
import java.util.Collections;
import java.util.List;
import java.util.Map;
import java.util.concurrent.locks.ReentrantLock;
import org.apache.hadoop.hbase.Cell;
import org.apache.hadoop.hbase.HConstants;
import org.apache.hadoop.hbase.KeyValue;
import org.apache.hadoop.hbase.client.Append;
import org.apache.hadoop.hbase.client.Increment;
import org.apache.hadoop.hbase.client.Result;
import org.apache.hadoop.hbase.util.Bytes;
import org.apache.phoenix.coprocessor.MetaDataProtocol;
import org.apache.phoenix.coprocessor.SequenceRegionObserver;
import org.apache.phoenix.exception.SQLExceptionCode;
import org.apache.phoenix.exception.SQLExceptionInfo;
import org.apache.phoenix.jdbc.PhoenixDatabaseMetaData;
import org.apache.phoenix.query.QueryConstants;
import org.apache.phoenix.schema.types.PBoolean;
import org.apache.phoenix.schema.types.PDataType;
import org.apache.phoenix.schema.types.PInteger;
import org.apache.phoenix.schema.types.PLong;
import org.apache.phoenix.util.ByteUtil;
import org.apache.phoenix.util.KeyValueUtil;
import org.apache.phoenix.util.SequenceUtil;
import com.google.common.collect.Lists;
import com.google.common.math.LongMath;
public class Sequence {
public static final int SUCCESS = 0;
public enum ValueOp {VALIDATE_SEQUENCE, INCREMENT_SEQUENCE};
public enum MetaOp {CREATE_SEQUENCE, DROP_SEQUENCE, RETURN_SEQUENCE};
// create empty Sequence key values used while created a sequence row
private static final KeyValue CURRENT_VALUE_KV = KeyValue.createFirstOnRow(ByteUtil.EMPTY_BYTE_ARRAY, SYSTEM_SEQUENCE_FAMILY_BYTES, CURRENT_VALUE_BYTES);
private static final KeyValue INCREMENT_BY_KV = KeyValue.createFirstOnRow(ByteUtil.EMPTY_BYTE_ARRAY, SYSTEM_SEQUENCE_FAMILY_BYTES, INCREMENT_BY_BYTES);
private static final KeyValue CACHE_SIZE_KV = KeyValue.createFirstOnRow(ByteUtil.EMPTY_BYTE_ARRAY, SYSTEM_SEQUENCE_FAMILY_BYTES, CACHE_SIZE_BYTES);
private static final KeyValue MIN_VALUE_KV = KeyValue.createFirstOnRow(ByteUtil.EMPTY_BYTE_ARRAY, SYSTEM_SEQUENCE_FAMILY_BYTES, MIN_VALUE_BYTES);
private static final KeyValue MAX_VALUE_KV = KeyValue.createFirstOnRow(ByteUtil.EMPTY_BYTE_ARRAY, SYSTEM_SEQUENCE_FAMILY_BYTES, MAX_VALUE_BYTES);
private static final KeyValue CYCLE_KV = KeyValue.createFirstOnRow(ByteUtil.EMPTY_BYTE_ARRAY, SYSTEM_SEQUENCE_FAMILY_BYTES, CYCLE_FLAG_BYTES);
private static final KeyValue LIMIT_REACHED_KV = KeyValue.createFirstOnRow(ByteUtil.EMPTY_BYTE_ARRAY, SYSTEM_SEQUENCE_FAMILY_BYTES, LIMIT_REACHED_FLAG_BYTES);
private static final List SEQUENCE_KV_COLUMNS = Arrays.asList(
CURRENT_VALUE_KV,
INCREMENT_BY_KV,
CACHE_SIZE_KV,
// The following three columns were added in 3.1/4.1
MIN_VALUE_KV,
MAX_VALUE_KV,
CYCLE_KV,
LIMIT_REACHED_KV
);
static {
Collections.sort(SEQUENCE_KV_COLUMNS, KeyValue.COMPARATOR);
}
// Pre-compute index of sequence key values to prevent binary search
private static final int CURRENT_VALUE_INDEX = SEQUENCE_KV_COLUMNS.indexOf(CURRENT_VALUE_KV);
private static final int INCREMENT_BY_INDEX = SEQUENCE_KV_COLUMNS.indexOf(INCREMENT_BY_KV);
private static final int CACHE_SIZE_INDEX = SEQUENCE_KV_COLUMNS.indexOf(CACHE_SIZE_KV);
private static final int MIN_VALUE_INDEX = SEQUENCE_KV_COLUMNS.indexOf(MIN_VALUE_KV);
private static final int MAX_VALUE_INDEX = SEQUENCE_KV_COLUMNS.indexOf(MAX_VALUE_KV);
private static final int CYCLE_INDEX = SEQUENCE_KV_COLUMNS.indexOf(CYCLE_KV);
private static final int LIMIT_REACHED_INDEX = SEQUENCE_KV_COLUMNS.indexOf(LIMIT_REACHED_KV);
public static final int NUM_SEQUENCE_KEY_VALUES = SEQUENCE_KV_COLUMNS.size();
private static final EmptySequenceCacheException EMPTY_SEQUENCE_CACHE_EXCEPTION = new EmptySequenceCacheException();
private final SequenceKey key;
private final ReentrantLock lock;
private List values;
public Sequence(SequenceKey key) {
if (key == null) throw new NullPointerException();
this.key = key;
this.lock = new ReentrantLock();
}
private void insertSequenceValue(SequenceValue value) {
if (values == null) {
values = Lists.newArrayListWithExpectedSize(1);
values.add(value);
} else {
int i = values.size()-1;
while (i >= 0 && values.get(i).timestamp > value.timestamp) {
i--;
}
// Don't insert another value if there's one at the same timestamp that is a delete
if (i >= 0 && values.get(i).timestamp == value.timestamp) {
if (values.get(i).isDeleted) {
throw new IllegalStateException("Unexpected delete marker at timestamp " + value.timestamp + " for "+ key);
}
values.set(i, value);
} else {
values.add(i+1, value);
}
}
}
private SequenceValue findSequenceValue(long timestamp) {
if (values == null) {
return null;
}
int i = values.size()-1;
while (i >= 0 && values.get(i).timestamp >= timestamp) {
i--;
}
if (i < 0) {
return null;
}
SequenceValue value = values.get(i);
return value.isDeleted ? null : value;
}
private long increment(SequenceValue value, ValueOp op, long numToAllocate) throws SQLException {
boolean increasingSeq = value.incrementBy > 0 && op != ValueOp.VALIDATE_SEQUENCE;
// check if the the sequence has already reached the min/max limit
if (value.limitReached && op != ValueOp.VALIDATE_SEQUENCE) {
if (value.cycle) {
value.limitReached=false;
throw EMPTY_SEQUENCE_CACHE_EXCEPTION;
} else {
SQLExceptionCode code =
increasingSeq ? SQLExceptionCode.SEQUENCE_VAL_REACHED_MAX_VALUE
: SQLExceptionCode.SEQUENCE_VAL_REACHED_MIN_VALUE;
throw SequenceUtil.getException(this.key.getSchemaName(),
this.key.getSequenceName(), code);
}
}
long returnValue = value.currentValue;
if (op == ValueOp.INCREMENT_SEQUENCE) {
boolean overflowOrUnderflow=false;
// advance currentValue while checking for overflow
try {
// advance by numToAllocate * the increment amount
value.currentValue = LongMath.checkedAdd(value.currentValue, numToAllocate * value.incrementBy);
} catch (ArithmeticException e) {
overflowOrUnderflow = true;
}
// set the limitReached flag (which will be checked the next time increment is called)
// if overflow or limit was reached
if (overflowOrUnderflow || (increasingSeq && value.currentValue > value.maxValue)
|| (!increasingSeq && value.currentValue < value.minValue)) {
value.limitReached=true;
}
}
return returnValue;
}
public long incrementValue(long timestamp, ValueOp op, long numToAllocate) throws SQLException {
SequenceValue value = findSequenceValue(timestamp);
if (value == null) {
throw EMPTY_SEQUENCE_CACHE_EXCEPTION;
}
if (isSequenceCacheExhausted(numToAllocate, value)) {
if (op == ValueOp.VALIDATE_SEQUENCE) {
return value.currentValue;
}
throw EMPTY_SEQUENCE_CACHE_EXCEPTION;
}
return increment(value, op, numToAllocate);
}
/**
* This method first checks whether value.currentValue = value.nextValue, this check is what
* determines whether we need to refresh the cache when evaluating NEXT VALUE FOR. Once
* current value reaches the next value we know the cache is exhausted as we give sequence
* values out one at time.
*
* However for bulk allocations, evaluated by NEXT VALUE FOR, we need a different check
* @see isSequenceCacheExhaustedForBulkAllocation
*
* Using the bulk allocation method for determining if the cache is exhausted for both cases
* works in most of the cases, however when dealing with CYCLEs and overflow and underflow, things
* break down due to things like sign changes that can happen if we overflow from a positive to
* a negative number and vice versa. Therefore, leaving both checks in place.
*
*/
private boolean isSequenceCacheExhausted(final long numToAllocate, final SequenceValue value) throws SQLException {
return value.currentValue == value.nextValue || (SequenceUtil.isBulkAllocation(numToAllocate) && isSequenceCacheExhaustedForBulkAllocation(numToAllocate, value));
}
/**
* This method checks whether there are sufficient values in the SequenceValue
* cached on the client to allocate the requested number of slots. It handles
* decreasing and increasing sequences as well as any overflows or underflows
* encountered.
*/
private boolean isSequenceCacheExhaustedForBulkAllocation(final long numToAllocate, final SequenceValue value) throws SQLException {
long targetSequenceValue;
performValidationForBulkAllocation(numToAllocate, value);
try {
targetSequenceValue = LongMath.checkedAdd(value.currentValue, numToAllocate * value.incrementBy);
} catch (ArithmeticException e) {
// Perform a CheckedAdd to make sure if over/underflow
// We don't treat this as the cache being exhausted as the current value may be valid in the case
// of no cycle, logic in increment() will take care of detecting we've hit the limit of the sequence
return false;
}
if (value.incrementBy > 0) {
return targetSequenceValue > value.nextValue;
} else {
return targetSequenceValue < value.nextValue;
}
}
/**
* @throws SQLException with the correct error code if sequence limit is reached with
* this request for allocation or we attempt to perform a bulk allocation on a sequence
* with cycles.
*/
private void performValidationForBulkAllocation(final long numToAllocate, final SequenceValue value)
throws SQLException {
boolean increasingSeq = value.incrementBy > 0 ? true : false;
// We don't support Bulk Allocations on sequences that have the CYCLE flag set to true
// Check for this here so we fail on expression evaluation and don't allow corner case
// whereby a client requests less than cached number of slots on sequence with cycle to succeed
if (value.cycle && !SequenceUtil.isCycleAllowed(numToAllocate)) {
throw new SQLExceptionInfo.Builder(SQLExceptionCode.NUM_SEQ_TO_ALLOCATE_NOT_SUPPORTED)
.setSchemaName(key.getSchemaName())
.setTableName(key.getSequenceName())
.build().buildException();
}
if (SequenceUtil.checkIfLimitReached(value.currentValue, value.minValue, value.maxValue, value.incrementBy, value.cacheSize, numToAllocate)) {
throw new SQLExceptionInfo.Builder(SequenceUtil.getLimitReachedErrorCode(increasingSeq))
.setSchemaName(key.getSchemaName())
.setTableName(key.getSequenceName())
.build().buildException();
}
}
public List newReturns() {
if (values == null) {
return Collections.emptyList();
}
List appends = Lists.newArrayListWithExpectedSize(values.size());
for (SequenceValue value : values) {
if (value.isInitialized() && value.currentValue != value.nextValue) {
appends.add(newReturn(value));
}
}
return appends;
}
public Append newReturn(long timestamp) throws EmptySequenceCacheException {
SequenceValue value = findSequenceValue(timestamp);
if (value == null) {
throw EMPTY_SEQUENCE_CACHE_EXCEPTION;
}
if (value.currentValue == value.nextValue) {
throw EMPTY_SEQUENCE_CACHE_EXCEPTION;
}
return newReturn(value);
}
private Append newReturn(SequenceValue value) {
byte[] key = this.key.getKey();
Append append = new Append(key);
byte[] opBuf = new byte[] {(byte)MetaOp.RETURN_SEQUENCE.ordinal()};
append.setAttribute(SequenceRegionObserver.OPERATION_ATTRIB, opBuf);
append.setAttribute(SequenceRegionObserver.CURRENT_VALUE_ATTRIB, PLong.INSTANCE.toBytes(value.nextValue));
Map> familyMap = append.getFamilyCellMap();
familyMap.put(PhoenixDatabaseMetaData.SYSTEM_SEQUENCE_FAMILY_BYTES, Arrays.asList(
(Cell)KeyValueUtil.newKeyValue(key, PhoenixDatabaseMetaData.SYSTEM_SEQUENCE_FAMILY_BYTES, PhoenixDatabaseMetaData.CURRENT_VALUE_BYTES, value.timestamp, PLong.INSTANCE.toBytes(value.currentValue)),
(Cell)KeyValueUtil.newKeyValue(key, PhoenixDatabaseMetaData.SYSTEM_SEQUENCE_FAMILY_BYTES, PhoenixDatabaseMetaData.LIMIT_REACHED_FLAG_BYTES, value.timestamp, PBoolean.INSTANCE.toBytes(value.limitReached))
));
return append;
}
public long currentValue(long timestamp) throws EmptySequenceCacheException {
SequenceValue value = findSequenceValue(timestamp);
if (value == null || value.isUnitialized()) {
throw EMPTY_SEQUENCE_CACHE_EXCEPTION;
}
return value.currentValue - value.incrementBy;
}
public ReentrantLock getLock() {
return lock;
}
public SequenceKey getKey() {
return key;
}
public long incrementValue(Result result, ValueOp op, long numToAllocate) throws SQLException {
// In this case, we don't definitely know the timestamp of the deleted sequence,
// but we know anything older is likely deleted. Worse case, we remove a sequence
// from the cache that we shouldn't have which will cause a gap in sequence values.
// In that case, we might get an error that a curr value was done on a sequence
// before a next val was. Not sure how to prevent that.
if (result.rawCells().length == 1) {
Cell errorKV = result.rawCells()[0];
int errorCode = PInteger.INSTANCE.getCodec().decodeInt(errorKV.getValueArray(), errorKV.getValueOffset(), SortOrder.getDefault());
SQLExceptionCode code = SQLExceptionCode.fromErrorCode(errorCode);
// TODO: We could have the server return the timestamps of the
// delete markers and we could insert them here, but this seems
// like overkill.
// if (code == SQLExceptionCode.SEQUENCE_UNDEFINED) {
// }
throw new SQLExceptionInfo.Builder(code)
.setSchemaName(key.getSchemaName())
.setTableName(key.getSequenceName())
.build().buildException();
}
// If we found the sequence, we update our cache with the new value
SequenceValue value = new SequenceValue(result, op, numToAllocate);
insertSequenceValue(value);
return increment(value, op, numToAllocate);
}
public Increment newIncrement(long timestamp, Sequence.ValueOp action, long numToAllocate) {
byte[] incKey = key.getKey();
byte[] incValue = Bytes.toBytes((long)action.ordinal());
Increment inc = new Increment(incKey);
// It doesn't matter what we set the amount too - we always use the values we get
// from the Get we do to prevent any race conditions. All columns that get added
// are returned with their current value
try {
inc.setTimeRange(MetaDataProtocol.MIN_TABLE_TIMESTAMP, timestamp);
inc.setAttribute(SequenceRegionObserver.NUM_TO_ALLOCATE, Bytes.toBytes(numToAllocate));
} catch (IOException e) {
throw new RuntimeException(e); // Impossible
}
for (KeyValue kv : SEQUENCE_KV_COLUMNS) {
try {
// Store the timestamp on the cell as well as HBase 1.2 seems to not
// be serializing over the time range (see HBASE-15698).
Cell cell = new KeyValue(incKey, 0, incKey.length,
kv.getFamilyArray(), kv.getFamilyOffset(), kv.getFamilyLength(),
kv.getQualifierArray(), kv.getQualifierOffset(), kv.getQualifierLength(),
timestamp,
KeyValue.Type.Put,
incValue, 0, incValue.length);
inc.add(cell);
} catch (IOException e) {
throw new RuntimeException(e); // Impossible
}
}
return inc;
}
/**
* Returns a KeyValue from the input result row
* @param kv an empty KeyValue used only to get the column family and column qualifier of the
* key value to be returned (if the sequence row is from a previous version)
* @param cellIndex index of the KeyValue to be returned (if the sequence row is from a previous version
* @return KeyValue
*/
private static KeyValue getKeyValue(Result r, KeyValue kv, int cellIndex) {
Cell[] cells = r.rawCells();
// if the sequence row is from a previous version then MIN_VALUE, MAX_VALUE, CYCLE and LIMIT_REACHED key values are not present,
// the sequence row has only three columns (INCREMENT_BY, CACHE_SIZE and CURRENT_VALUE) and the order of the cells
// in the array returned by rawCells() is not what what we expect so use getColumnLatestCell() to get the cell we want
Cell cell = cells.length == NUM_SEQUENCE_KEY_VALUES ? cells[cellIndex] :
r.getColumnLatestCell(kv.getFamilyArray(), kv.getFamilyOffset(), kv.getFamilyLength(), kv.getQualifierArray(), kv.getQualifierOffset(), kv.getQualifierLength());
return org.apache.hadoop.hbase.KeyValueUtil.ensureKeyValue(cell);
}
private static KeyValue getKeyValue(Result r, KeyValue kv) {
return getKeyValue(r, kv, SEQUENCE_KV_COLUMNS.indexOf(kv));
}
public static KeyValue getCurrentValueKV(Result r) {
return getKeyValue(r, CURRENT_VALUE_KV, CURRENT_VALUE_INDEX);
}
public static KeyValue getIncrementByKV(Result r) {
return getKeyValue(r, INCREMENT_BY_KV, INCREMENT_BY_INDEX);
}
public static KeyValue getCacheSizeKV(Result r) {
return getKeyValue(r, CACHE_SIZE_KV, CACHE_SIZE_INDEX);
}
public static KeyValue getMinValueKV(Result r) {
return getKeyValue(r, MIN_VALUE_KV, MIN_VALUE_INDEX);
}
public static KeyValue getMaxValueKV(Result r) {
return getKeyValue(r, MAX_VALUE_KV, MAX_VALUE_INDEX);
}
public static KeyValue getCycleKV(Result r) {
return getKeyValue(r, CYCLE_KV, CYCLE_INDEX);
}
public static KeyValue getLimitReachedKV(Result r) {
return getKeyValue(r, LIMIT_REACHED_KV, LIMIT_REACHED_INDEX);
}
public static void replaceCurrentValueKV(List kvs, KeyValue currentValueKV) {
kvs.set(CURRENT_VALUE_INDEX, currentValueKV);
}
public static void replaceMinValueKV(List kvs, KeyValue minValueKV) {
kvs.set(MIN_VALUE_INDEX, minValueKV);
}
public static void replaceMaxValueKV(List kvs, KeyValue maxValueKV) {
kvs.set(MAX_VALUE_INDEX, maxValueKV);
}
public static void replaceCycleValueKV(List kvs, KeyValue cycleValueKV) {
kvs.set(CYCLE_INDEX, cycleValueKV);
}
public static void replaceLimitReachedKV(List kvs, KeyValue limitReachedKV) {
kvs.set(LIMIT_REACHED_INDEX, limitReachedKV);
}
/**
* Returns the KeyValues of r if it contains the expected number of KeyValues,
* else returns a list of KeyValues corresponding to SEQUENCE_KV_COLUMNS
*/
public static List getCells(Result r, int numKVs) {
// if the sequence row is from a previous version
if (r.rawCells().length == numKVs )
return Lists.newArrayList(r.rawCells());
// else we need to handle missing MIN_VALUE, MAX_VALUE, CYCLE and LIMIT_REACHED KeyValues
List cellList = Lists.newArrayListWithCapacity(NUM_SEQUENCE_KEY_VALUES);
for (KeyValue kv : SEQUENCE_KV_COLUMNS) {
cellList.add(getKeyValue(r,kv));
}
return cellList;
}
private static final class SequenceValue {
public final long incrementBy;
public final long timestamp;
public final long cacheSize;
public long currentValue;
public long nextValue;
public long minValue;
public long maxValue;
public boolean cycle;
public boolean isDeleted;
public boolean limitReached;
public SequenceValue(long timestamp, long minValue, long maxValue, boolean cycle) {
this(timestamp, false);
this.minValue = minValue;
this.maxValue = maxValue;
this.cycle = cycle;
}
public SequenceValue(long timestamp, boolean isDeleted) {
this.timestamp = timestamp;
this.isDeleted = isDeleted;
this.incrementBy = 0;
this.limitReached = false;
this.cacheSize = 0;
}
public boolean isInitialized() {
return this.incrementBy != 0;
}
public boolean isUnitialized() {
return this.incrementBy == 0;
}
public SequenceValue(Result r, ValueOp op, long numToAllocate) {
KeyValue currentValueKV = getCurrentValueKV(r);
KeyValue incrementByKV = getIncrementByKV(r);
KeyValue cacheSizeKV = getCacheSizeKV(r);
KeyValue minValueKV = getMinValueKV(r);
KeyValue maxValueKV = getMaxValueKV(r);
KeyValue cycleKV = getCycleKV(r);
this.timestamp = currentValueKV.getTimestamp();
this.nextValue = PLong.INSTANCE.getCodec().decodeLong(currentValueKV.getValueArray(), currentValueKV.getValueOffset(), SortOrder.getDefault());
this.incrementBy = PLong.INSTANCE.getCodec().decodeLong(incrementByKV.getValueArray(), incrementByKV.getValueOffset(), SortOrder.getDefault());
this.cacheSize = PLong.INSTANCE.getCodec().decodeLong(cacheSizeKV.getValueArray(), cacheSizeKV.getValueOffset(), SortOrder.getDefault());
this.minValue = PLong.INSTANCE.getCodec().decodeLong(minValueKV.getValueArray(), minValueKV.getValueOffset(), SortOrder.getDefault());
this.maxValue = PLong.INSTANCE.getCodec().decodeLong(maxValueKV.getValueArray(), maxValueKV.getValueOffset(), SortOrder.getDefault());
this.cycle = (Boolean) PBoolean.INSTANCE.toObject(cycleKV.getValueArray(), cycleKV.getValueOffset(), cycleKV.getValueLength());
this.limitReached = false;
currentValue = nextValue;
if (op != ValueOp.VALIDATE_SEQUENCE) {
// We can't just take the max of numToAllocate and cacheSize
// We need to handle a valid edgecase where a client requests bulk allocation of
// a number of slots that are less than cache size of the sequence
currentValue -= incrementBy * (SequenceUtil.isBulkAllocation(numToAllocate) ? numToAllocate : cacheSize);
}
}
}
public boolean returnValue(Result result) throws SQLException {
Cell statusKV = result.rawCells()[0];
if (statusKV.getValueLength() == 0) { // No error, but unable to return sequence values
return false;
}
long timestamp = statusKV.getTimestamp();
int statusCode = PInteger.INSTANCE.getCodec().decodeInt(statusKV.getValueArray(), statusKV.getValueOffset(), SortOrder.getDefault());
if (statusCode == SUCCESS) { // Success - update nextValue down to currentValue
SequenceValue value = findSequenceValue(timestamp);
if (value == null) {
throw new EmptySequenceCacheException(key.getSchemaName(),key.getSequenceName());
}
return true;
}
SQLExceptionCode code = SQLExceptionCode.fromErrorCode(statusCode);
// TODO: We could have the server return the timestamps of the
// delete markers and we could insert them here, but this seems
// like overkill.
// if (code == SQLExceptionCode.SEQUENCE_UNDEFINED) {
// }
throw new SQLExceptionInfo.Builder(code)
.setSchemaName(key.getSchemaName())
.setTableName(key.getSequenceName())
.build().buildException();
}
public Append createSequence(long startWith, long incrementBy, long cacheSize, long timestamp, long minValue, long maxValue, boolean cycle) {
byte[] key = this.key.getKey();
Append append = new Append(key);
append.setAttribute(SequenceRegionObserver.OPERATION_ATTRIB, new byte[] {(byte)MetaOp.CREATE_SEQUENCE.ordinal()});
if (timestamp != HConstants.LATEST_TIMESTAMP) {
append.setAttribute(SequenceRegionObserver.MAX_TIMERANGE_ATTRIB, Bytes.toBytes(timestamp));
}
Map> familyMap = append.getFamilyCellMap();
byte[] startWithBuf = PLong.INSTANCE.toBytes(startWith);
familyMap.put(PhoenixDatabaseMetaData.SYSTEM_SEQUENCE_FAMILY_BYTES, Arrays.asList(
KeyValueUtil.newKeyValue(key, PhoenixDatabaseMetaData.SYSTEM_SEQUENCE_FAMILY_BYTES, QueryConstants.EMPTY_COLUMN_BYTES, timestamp, ByteUtil.EMPTY_BYTE_ARRAY),
KeyValueUtil.newKeyValue(key, PhoenixDatabaseMetaData.SYSTEM_SEQUENCE_FAMILY_BYTES, PhoenixDatabaseMetaData.CURRENT_VALUE_BYTES, timestamp, startWithBuf),
KeyValueUtil.newKeyValue(key, PhoenixDatabaseMetaData.SYSTEM_SEQUENCE_FAMILY_BYTES, PhoenixDatabaseMetaData.START_WITH_BYTES, timestamp, startWithBuf),
KeyValueUtil.newKeyValue(key, PhoenixDatabaseMetaData.SYSTEM_SEQUENCE_FAMILY_BYTES, PhoenixDatabaseMetaData.INCREMENT_BY_BYTES, timestamp, PLong.INSTANCE.toBytes(incrementBy)),
KeyValueUtil.newKeyValue(key, PhoenixDatabaseMetaData.SYSTEM_SEQUENCE_FAMILY_BYTES, PhoenixDatabaseMetaData.CACHE_SIZE_BYTES, timestamp, PLong.INSTANCE.toBytes(cacheSize)),
KeyValueUtil.newKeyValue(key, PhoenixDatabaseMetaData.SYSTEM_SEQUENCE_FAMILY_BYTES, PhoenixDatabaseMetaData.MIN_VALUE_BYTES, timestamp, PLong.INSTANCE.toBytes(minValue)),
KeyValueUtil.newKeyValue(key, PhoenixDatabaseMetaData.SYSTEM_SEQUENCE_FAMILY_BYTES, PhoenixDatabaseMetaData.MAX_VALUE_BYTES, timestamp, PLong.INSTANCE.toBytes(maxValue)),
KeyValueUtil.newKeyValue(key, PhoenixDatabaseMetaData.SYSTEM_SEQUENCE_FAMILY_BYTES, PhoenixDatabaseMetaData.CYCLE_FLAG_BYTES, timestamp, PBoolean.INSTANCE.toBytes(cycle)),
KeyValueUtil.newKeyValue(key, PhoenixDatabaseMetaData.SYSTEM_SEQUENCE_FAMILY_BYTES, PhoenixDatabaseMetaData.LIMIT_REACHED_FLAG_BYTES, timestamp, PDataType.FALSE_BYTES)
));
return append;
}
public long createSequence(Result result, long minValue, long maxValue, boolean cycle) throws SQLException {
Cell statusKV = result.rawCells()[0];
long timestamp = statusKV.getTimestamp();
int statusCode = PInteger.INSTANCE.getCodec().decodeInt(statusKV.getValueArray(), statusKV.getValueOffset(), SortOrder.getDefault());
if (statusCode == 0) { // Success - add sequence value and return timestamp
SequenceValue value = new SequenceValue(timestamp, minValue, maxValue, cycle);
insertSequenceValue(value);
return timestamp;
}
SQLExceptionCode code = SQLExceptionCode.fromErrorCode(statusCode);
throw new SQLExceptionInfo.Builder(code)
.setSchemaName(key.getSchemaName())
.setTableName(key.getSequenceName())
.build().buildException();
}
public Append dropSequence(long timestamp) {
byte[] key = this.key.getKey();
Append append = new Append(key);
append.setAttribute(SequenceRegionObserver.OPERATION_ATTRIB, new byte[] {(byte)MetaOp.DROP_SEQUENCE.ordinal()});
if (timestamp != HConstants.LATEST_TIMESTAMP) {
append.setAttribute(SequenceRegionObserver.MAX_TIMERANGE_ATTRIB, Bytes.toBytes(timestamp));
}
Map> familyMap = append.getFamilyCellMap();
familyMap.put(PhoenixDatabaseMetaData.SYSTEM_SEQUENCE_FAMILY_BYTES, Arrays.| asList(
(Cell)KeyValueUtil.newKeyValue(key, PhoenixDatabaseMetaData.SYSTEM_SEQUENCE_FAMILY_BYTES, QueryConstants.EMPTY_COLUMN_BYTES, timestamp, ByteUtil.EMPTY_BYTE_ARRAY)));
return append;
}
public long dropSequence(Result result) throws SQLException {
Cell statusKV = result.rawCells()[0];
long timestamp = statusKV.getTimestamp();
int statusCode = PInteger.INSTANCE.getCodec().decodeInt(statusKV.getValueArray(), statusKV.getValueOffset(), SortOrder.getDefault());
SQLExceptionCode code = statusCode == 0 ? null : SQLExceptionCode.fromErrorCode(statusCode);
if (code == null) {
// Insert delete marker so that point-in-time sequences work
insertSequenceValue(new SequenceValue(timestamp, true));
return timestamp;
}
// TODO: We could have the server return the timestamps of the
// delete markers and we could insert them here, but this seems
// like overkill.
// if (code == SQLExceptionCode.SEQUENCE_UNDEFINED) {
// }
throw new SQLExceptionInfo.Builder(code)
.setSchemaName(key.getSchemaName())
.setTableName(key.getSequenceName())
.build().buildException();
}
public static String getCreateTableStatement(String schema, int nSaltBuckets) {
if (nSaltBuckets <= 0) {
return schema;
}
return schema + "," + PhoenixDatabaseMetaData.SALT_BUCKETS + "=" + nSaltBuckets;
}
}
| | | | | | | | | | © 2015 - 2024 Weber Informatics LLC | Privacy Policy