alex-ninja commented on a change in pull request #1117: URL: https://github.com/apache/cassandra/pull/1117#discussion_r705982846
########## File path: src/java/org/apache/cassandra/db/virtual/AbstractMutableVirtualTable.java ########## @@ -0,0 +1,198 @@ +/* + * 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.virtual; + +import java.nio.ByteBuffer; + +import com.google.common.collect.BoundType; +import com.google.common.collect.Range; + +import org.apache.cassandra.db.ClusteringBound; +import org.apache.cassandra.db.ClusteringPrefix; +import org.apache.cassandra.db.DecoratedKey; +import org.apache.cassandra.db.Slice; +import org.apache.cassandra.db.marshal.CompositeType; +import org.apache.cassandra.db.partitions.PartitionUpdate; +import org.apache.cassandra.db.rows.Cell; +import org.apache.cassandra.exceptions.InvalidRequestException; +import org.apache.cassandra.schema.TableMetadata; + +/** + * An abstract virtual table implementation that builds the resultset on demand and allows fine-grained source + * modification via INSERT/UPDATE, DELETE and TRUNCATE operations. + */ +public abstract class AbstractMutableVirtualTable extends AbstractVirtualTable +{ + + protected AbstractMutableVirtualTable(TableMetadata metadata) + { + super(metadata); + } + + @Override + public void apply(PartitionUpdate update) + { + Object[] partitionKeyColumnValues = extractPartitionKeyColumnValues(update.partitionKey()); + + if (update.deletionInfo().isLive()) + update.forEach(row -> + { + Comparable<?>[] clusteringColumnValues = extractClusteringColumnValues(row.clustering()); + + if (row.deletion().isLive()) + { + row.forEach(columnData -> + { + if (columnData.column().isComplex()) + throw new InvalidRequestException("Complex type columns are not supported by table " + metadata); + + Cell<?> cell = (Cell<?>) columnData; + String columnName = extractColumnName(cell); + + if (cell.isTombstone()) + applyColumnDeletion(partitionKeyColumnValues, clusteringColumnValues, columnName); + else + applyColumnUpdate(partitionKeyColumnValues, + clusteringColumnValues, + columnName, + extractColumnValue(cell)); + }); + } + else + applyRowDeletion(partitionKeyColumnValues, clusteringColumnValues); + }); + else + { + // MutableDeletionInfo may have partition delete or range tombstone list or both + if (update.deletionInfo().hasRanges()) + update.deletionInfo() + .rangeIterator(false) + .forEachRemaining(rt -> convertAndApplyRangeTombstone(partitionKeyColumnValues, rt.deletedSlice())); + + if (!update.deletionInfo().getPartitionDeletion().isLive()) + applyPartitionDeletion(partitionKeyColumnValues); + } + } + + protected void applyPartitionDeletion(Object[] partitionKeyColumnValues) + { + throw new InvalidRequestException("Partition deletion is not supported by table " + metadata); + + } + + private void convertAndApplyRangeTombstone(Object[] partitionKeyColumnValues, Slice slice) + { + ClusteringBound<?> startClusteringColumns = slice.start(); + Comparable<?>[] startClusteringColumnValues = extractClusteringColumnValues(startClusteringColumns); + BoundType startClusteringColumnBoundType = startClusteringColumns.isInclusive() ? BoundType.CLOSED : BoundType.OPEN; + + ClusteringBound<?> endClusteringColumns = slice.end(); + Comparable<?>[] endClusteringColumnValues = extractClusteringColumnValues(endClusteringColumns); + BoundType endClusteringColumnBoundType = endClusteringColumns.isInclusive() ? BoundType.CLOSED : BoundType.OPEN; + + int clusteringColumnsPrefixLength = Math.max(startClusteringColumnValues.length, endClusteringColumnValues.length) - 1; + Comparable<?>[] clusteringColumnValuesPrefix = new Comparable<?>[clusteringColumnsPrefixLength]; + System.arraycopy(startClusteringColumnValues, 0, clusteringColumnValuesPrefix, 0, clusteringColumnsPrefixLength); + + Range<Comparable<?>> range; + if (startClusteringColumnValues.length < endClusteringColumnValues.length) + range = Range.upTo(endClusteringColumnValues[endClusteringColumnValues.length - 1], endClusteringColumnBoundType); + else if (startClusteringColumnValues.length > endClusteringColumnValues.length) + range = Range.downTo(startClusteringColumnValues[startClusteringColumnValues.length - 1], startClusteringColumnBoundType); + else + range = Range.range(startClusteringColumnValues[startClusteringColumnValues.length - 1], startClusteringColumnBoundType, + endClusteringColumnValues[endClusteringColumnValues.length - 1], endClusteringColumnBoundType); + + applyRangeTombstone(partitionKeyColumnValues, clusteringColumnValuesPrefix, range); + } + + /** + * This method accepts parsed parts of a corresponding range tombstone. There is a tricky logic for + * {@code clusteringColumnValuesPrefix} calculation. It consists of clustering columns that have equality condition. + * It is worth mentioning that it is only possible to have clusering columns from first levels. For example, if + * there are three clustering columns specified in the range combstone: {@code c1='c1_1' AND c2='c2_1' AND c3>'c3_1'}, + * then {@code clusteringColumnValuesPrefix} will have ["c1_1", "c2_1"] values. In case of a single clustering column + * the prefix will be empty. + * + * @param partitionKeyColumnValues is a non-empty array of partition key columns + * @param clusteringColumnValuesPrefix is an array (it may be empty!) of clustering columns with equality condition + * @param range is a range of values for the last clustering column + */ + protected void applyRangeTombstone(Object[] partitionKeyColumnValues, + Comparable<?>[] clusteringColumnValuesPrefix, + Range<Comparable<?>> range) + { + throw new InvalidRequestException("Range deletion is not supported by table " + metadata); + } + + protected void applyRowDeletion(Object[] partitionKeyColumnValues, Comparable<?>[] clusteringColumnValues) + { + throw new InvalidRequestException("Row deletion is not supported by table " + metadata); + } + + protected void applyColumnDeletion(Object[] partitionKeyColumnValues, Comparable<?>[] clusteringColumnValues, String columnName) + { + throw new InvalidRequestException("Column deletion is not supported by table " + metadata); + } + + protected void applyColumnUpdate(Object[] partitionKeyColumnValues, Comparable<?>[] clusteringColumnValues, + String columnName, Object columnValue) + { + throw new InvalidRequestException("Column modification is not supported by table " + metadata); + } + + private Object[] extractPartitionKeyColumnValues(DecoratedKey partitionKey) + { + if (metadata.partitionKeyType instanceof CompositeType) + { + ByteBuffer[] partitionKeyColumnBytes = ((CompositeType) metadata.partitionKeyType).split(partitionKey.getKey()); + Object[] partitionKeyColumnValues = new Object[partitionKeyColumnBytes.length]; + for (int i = 0; i < partitionKeyColumnValues.length; i++) + { + partitionKeyColumnValues[i] = metadata.partitionKeyColumns().get(i).type.compose(partitionKeyColumnBytes[i]); + } + return partitionKeyColumnValues; + + } + else + return new Object[]{metadata.partitionKeyType.compose(partitionKey.getKey())}; + } + + private Comparable<?>[] extractClusteringColumnValues(ClusteringPrefix<?> clusteringColumns) + { + // clusteringColumns.size() may be less than metadata.clusteringColumns().size() since not all clustering + // columns have to be always specified + Comparable<?>[] clusteringColumnValues = new Comparable<?>[clusteringColumns.size()]; + for (int i = 0; i < clusteringColumnValues.length; i++) + { + // VTs do not support UDTs, so casting is safe here because clustering columns of simple types are always comparable Review comment: Ok, makes sense. Done. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
