/*
* 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.cassandra.db;
import java.io.DataInput;
import java.io.IOError;
import java.io.IOException;
import java.net.InetAddress;
import java.nio.ByteBuffer;
import java.security.MessageDigest;
import java.util.ArrayList;
import java.util.Iterator;
import java.util.List;
import com.google.common.collect.AbstractIterator;
import org.apache.cassandra.config.CFMetaData;
import org.apache.cassandra.cql3.CFDefinition;
import org.apache.cassandra.db.marshal.*;
import org.apache.cassandra.io.sstable.Descriptor;
import org.apache.cassandra.io.util.DataOutputBuffer;
import org.apache.cassandra.serializers.MarshalException;
import org.apache.cassandra.utils.Allocator;
import org.apache.cassandra.utils.ByteBufferUtil;
import org.apache.cassandra.utils.FBUtilities;
import org.apache.cassandra.utils.HeapAllocator;
/**
* Column is immutable, which prevents all kinds of confusion in a multithreaded environment.
*/
public class Column implements OnDiskAtom
{
public static final int MAX_NAME_LENGTH = FBUtilities.MAX_UNSIGNED_SHORT;
public static final ColumnSerializer serializer = new ColumnSerializer();
public static OnDiskAtom.Serializer onDiskSerializer()
{
return OnDiskAtom.Serializer.instance;
}
/**
* For 2.0-formatted sstables (where column count is not stored), @param count should be Integer.MAX_VALUE,
* and we will look for the end-of-row column name marker instead of relying on that.
*/
public static Iterator<OnDiskAtom> onDiskIterator(final DataInput in, final int count, final ColumnSerializer.Flag flag, final int expireBefore, final Descriptor.Version version)
{
return new AbstractIterator<OnDiskAtom>()
{
int i = 0;
protected OnDiskAtom computeNext()
{
if (i++ >= count)
return endOfData();
OnDiskAtom atom;
try
{
atom = onDiskSerializer().deserializeFromSSTable(in, flag, expireBefore, version);
}
catch (IOException e)
{
throw new IOError(e);
}
if (atom == null)
return endOfData();
return atom;
}
};
}
protected final ByteBuffer name;
protected final ByteBuffer value;
protected final long timestamp;
Column(ByteBuffer name)
{
this(name, ByteBufferUtil.EMPTY_BYTE_BUFFER);
}
public Column(ByteBuffer name, ByteBuffer value)
{
this(name, value, 0);
}
public Column(ByteBuffer name, ByteBuffer value, long timestamp)
{
assert name != null;
assert value != null;
assert name.remaining() <= Column.MAX_NAME_LENGTH;
this.name = name;
this.value = value;
this.timestamp = timestamp;
}
public Column withUpdatedName(ByteBuffer newName)
{
return new Column(newName, value, timestamp);
}
public Column withUpdatedTimestamp(long newTimestamp)
{
return new Column(name, value, newTimestamp);
}
public ByteBuffer name()
{
return name;
}
public ByteBuffer value()
{
return value;
}
public long timestamp()
{
return timestamp;
}
public long minTimestamp()
{
return timestamp;
}
public long maxTimestamp()
{
return timestamp;
}
public boolean isMarkedForDelete(long now)
{
return false;
}
public boolean isLive(long now)
{
return !isMarkedForDelete(now);
}
// Don't call unless the column is actually marked for delete.
public long getMarkedForDeleteAt()
{
return Long.MAX_VALUE;
}
public int dataSize()
{
return name().remaining() + value.remaining() + TypeSizes.NATIVE.sizeof(timestamp);
}
public int serializedSize(TypeSizes typeSizes)
{
/*
* Size of a column is =
* size of a name (short + length of the string)
* + 1 byte to indicate if the column has been deleted
* + 8 bytes for timestamp
* + 4 bytes which basically indicates the size of the byte array
* + entire byte array.
*/
int nameSize = name.remaining();
int valueSize = value.remaining();
return typeSizes.sizeof((short) nameSize) + nameSize + 1 + typeSizes.sizeof(timestamp) + typeSizes.sizeof(valueSize) + valueSize;
}
public long serializedSizeForSSTable()
{
return serializedSize(TypeSizes.NATIVE);
}
public int serializationFlags()
{
return 0;
}
public Column diff(Column column)
{
if (timestamp() < column.timestamp())
return column;
return null;
}
public void updateDigest(MessageDigest digest)
{
digest.update(name.duplicate());
digest.update(value.duplicate());
DataOutputBuffer buffer = new DataOutputBuffer();
try
{
buffer.writeLong(timestamp);
buffer.writeByte(serializationFlags());
}
catch (IOException e)
{
throw new RuntimeException(e);
}
digest.update(buffer.getData(), 0, buffer.getLength());
}
public int getLocalDeletionTime()
{
return Integer.MAX_VALUE;
}
public Column reconcile(Column column)
{
return reconcile(column, HeapAllocator.instance);
}
public Column reconcile(Column column, Allocator allocator)
{
// tombstones take precedence. (if both are tombstones, then it doesn't matter which one we use.)
if (isMarkedForDelete(System.currentTimeMillis()))
return timestamp() < column.timestamp() ? column : this;
if (column.isMarkedForDelete(System.currentTimeMillis()))
return timestamp() > column.timestamp() ? this : column;
// break ties by comparing values.
if (timestamp() == column.timestamp())
return value().compareTo(column.value()) < 0 ? column : this;
// neither is tombstoned and timestamps are different
return timestamp() < column.timestamp() ? column : this;
}
@Override
public boolean equals(Object o)
{
if (this == o)
return true;
if (o == null || getClass() != o.getClass())
return false;
Column column = (Column)o;
if (timestamp != column.timestamp)
return false;
if (!name.equals(column.name))
return false;
return value.equals(column.value);
}
@Override
public int hashCode()
{
int result = name != null ? name.hashCode() : 0;
result = 31 * result + (value != null ? value.hashCode() : 0);
result = 31 * result + (int)(timestamp ^ (timestamp >>> 32));
return result;
}
public Column localCopy(ColumnFamilyStore cfs)
{
return localCopy(cfs, HeapAllocator.instance);
}
public Column localCopy(ColumnFamilyStore cfs, Allocator allocator)
{
return new Column(cfs.internOrCopy(name, allocator), allocator.clone(value), timestamp);
}
public String getString(AbstractType<?> comparator)
{
StringBuilder sb = new StringBuilder();
sb.append(comparator.getString(name));
sb.append(":");
sb.append(isMarkedForDelete(System.currentTimeMillis()));
sb.append(":");
sb.append(value.remaining());
sb.append("@");
sb.append(timestamp());
return sb.toString();
}
protected void validateName(CFMetaData metadata) throws MarshalException
{
metadata.comparator.validate(name());
}
public void validateFields(CFMetaData metadata) throws MarshalException
{
validateName(metadata);
CFDefinition cfdef = metadata.getCfDef();
// If this is a CQL table, we need to pull out the CQL column name to look up the correct column type.
// (Note that COMPACT composites are handled by validateName, above.)
ByteBuffer internalName = (cfdef.isComposite && !cfdef.isCompact)
? ((CompositeType) metadata.comparator).extractLastComponent(name)
: name;
AbstractType<?> valueValidator = metadata.getValueValidator(internalName);
if (valueValidator != null)
valueValidator.validate(value());
}
public boolean hasIrrelevantData(int gcBefore)
{
return getLocalDeletionTime() < gcBefore;
}
public static Column create(ByteBuffer name, ByteBuffer value, long timestamp, int ttl, CFMetaData metadata)
{
if (ttl <= 0)
ttl = metadata.getDefaultTimeToLive();
return ttl > 0
? new ExpiringColumn(name, value, timestamp, ttl)
: new Column(name, value, timestamp);
}
public static Column create(String value, long timestamp, String... names)
{
return new Column(decomposeName(names), UTF8Type.instance.decompose(value), timestamp);
}
public static Column create(int value, long timestamp, String... names)
{
return new Column(decomposeName(names), Int32Type.instance.decompose(value), timestamp);
}
public static Column create(boolean value, long timestamp, String... names)
{
return new Column(decomposeName(names), BooleanType.instance.decompose(value), timestamp);
}
public static Column create(double value, long timestamp, String... names)
{
return new Column(decomposeName(names), DoubleType.instance.decompose(value), timestamp);
}
public static Column create(ByteBuffer value, long timestamp, String... names)
{
return new Column(decomposeName(names), value, timestamp);
}
public static Column create(InetAddress value, long timestamp, String... names)
{
return new Column(decomposeName(names), InetAddressType.instance.decompose(value), timestamp);
}
static ByteBuffer decomposeName(String... names)
{
assert names.length > 0;
if (names.length == 1)
return UTF8Type.instance.decompose(names[0]);
// not super performant. at this time, only infrequently called schema code uses this.
List<AbstractType<?>> types = new ArrayList<AbstractType<?>>(names.length);
for (int i = 0; i < names.length; i++)
types.add(UTF8Type.instance);
CompositeType.Builder builder = new CompositeType.Builder(CompositeType.getInstance(types));
for (String name : names)
builder.add(UTF8Type.instance.decompose(name));
return builder.build();
}
}