io.deephaven.engine.table.impl.updateby.UpdateByOperator Maven / Gradle / Ivy
Go to download
Show more of this group Show more artifacts with this name
Show all versions of deephaven-engine-table Show documentation
Show all versions of deephaven-engine-table Show documentation
Engine Table: Implementation and closely-coupled utilities
package io.deephaven.engine.table.impl.updateby;
import io.deephaven.chunk.Chunk;
import io.deephaven.chunk.IntChunk;
import io.deephaven.chunk.LongChunk;
import io.deephaven.chunk.attributes.Values;
import io.deephaven.engine.rowset.RowSequence;
import io.deephaven.engine.rowset.RowSet;
import io.deephaven.engine.rowset.chunkattributes.OrderedRowKeys;
import io.deephaven.engine.table.ColumnSource;
import io.deephaven.engine.table.ModifiedColumnSet;
import io.deephaven.engine.table.Table;
import io.deephaven.engine.table.TableUpdate;
import io.deephaven.engine.table.impl.MatchPair;
import io.deephaven.engine.table.impl.QueryTable;
import io.deephaven.engine.table.impl.util.RowRedirection;
import io.deephaven.util.SafeCloseable;
import org.jetbrains.annotations.NotNull;
import org.jetbrains.annotations.Nullable;
import javax.annotation.OverridingMethodsMustInvokeSuper;
import java.util.Map;
/**
* An operator that performs a specific computation for {@link Table#updateBy}. When adding implementations of this
* interface, the pattern of calls will be as follows.
*
*
* - {@link UpdateByOperator#initializeCumulative(Context, long, long, RowSet)} for cumulative operators or
* {@link UpdateByOperator#initializeRolling(Context, RowSet)} (Context)} for windowed operators
* - {@link UpdateByOperator.Context#accumulateCumulative(RowSequence, Chunk[], LongChunk, int)} for cumulative
* operators or
* {@link UpdateByOperator.Context#accumulateRolling(RowSequence, Chunk[], LongChunk, LongChunk, IntChunk, IntChunk, int)}
* for windowed operators
* - {@link #finishUpdate(UpdateByOperator.Context)}
*
*/
public abstract class UpdateByOperator {
protected final MatchPair pair;
protected final String[] affectingColumns;
protected final long reverseWindowScaleUnits;
protected final long forwardWindowScaleUnits;
protected final String timestampColumnName;
protected final boolean isWindowed;
protected RowRedirection rowRedirection;
/**
* The input modifiedColumnSet for this operator
*/
protected ModifiedColumnSet inputModifiedColumnSet;
/**
* The output modifiedColumnSet for this operator
*/
protected ModifiedColumnSet outputModifiedColumnSet;
/**
* A context item for use with updateBy operators
*/
public abstract static class Context implements SafeCloseable {
protected int nullCount = 0;
protected LongChunk affectedPosChunk;
protected LongChunk influencerPosChunk;
public boolean isValueValid(long atKey) {
throw new UnsupportedOperationException(
"isValueValid() must be overridden by time-aware cumulative operators");
}
protected abstract void setValueChunks(@NotNull Chunk[] valueChunks);
protected void setPosChunks(final LongChunk affectedPosChunk,
final LongChunk influencerPosChunk) {
this.affectedPosChunk = affectedPosChunk;
this.influencerPosChunk = influencerPosChunk;
}
/**
* Add values to the operators current data set
*
* @param pos the index in the associated chunk where this value can be found. Depending on the usage, might be
* a values chunk (for cumulative operators) or an influencer values chunk (for windowed). It is the task
* of the operator to pull the data from the chunk and use it properly
* @param count the number of items to push from the chunk
*/
protected abstract void push(int pos, int count);
/**
* Remove values from the operators current data set. This is only valid for windowed operators as cumulative
* operators only append values
*
* @param count the number of items to pop from the data set
*/
protected void pop(int count) {
throw new UnsupportedOperationException("pop() must be overriden by rolling operators");
}
public abstract void accumulateCumulative(RowSequence inputKeys,
Chunk[] valueChunkArr,
LongChunk tsChunk,
int len);
public abstract void accumulateRolling(RowSequence inputKeys,
Chunk[] influencerValueChunkArr,
LongChunk affectedPosChunk,
LongChunk influencerPosChunk,
IntChunk pushChunk,
IntChunk popChunk,
int len);
/**
* Write the current value for this row to the output chunk
*/
protected abstract void writeToOutputChunk(int outIdx);
/**
* Write the output chunk to the output column
*/
protected abstract void writeToOutputColumn(@NotNull final RowSequence inputKeys);
/**
* Reset the operator data values to a known state. This may occur during initialization or when a windowed
* operator has an empty window
*/
@OverridingMethodsMustInvokeSuper
protected abstract void reset();
}
protected UpdateByOperator(@NotNull final MatchPair pair,
@NotNull final String[] affectingColumns,
@Nullable final String timestampColumnName,
final long reverseWindowScaleUnits,
final long forwardWindowScaleUnits,
final boolean isWindowed) {
this.pair = pair;
this.affectingColumns = affectingColumns;
this.timestampColumnName = timestampColumnName;
this.reverseWindowScaleUnits = reverseWindowScaleUnits;
this.forwardWindowScaleUnits = forwardWindowScaleUnits;
this.isWindowed = isWindowed;
}
/**
* Create an uninitialized copy of this operator. {@link #initializeSources(Table, RowRedirection)} must be called
* before this operator can be used.
*
* @return a copy of this operator
*/
public abstract UpdateByOperator copy();
/**
* Initialize this operator with a specific source table (and row redirection if needed). This will be called
* exactly once per operator.
*/
public abstract void initializeSources(@NotNull Table source, @Nullable RowRedirection rowRedirection);
/**
* Initialize the bucket context for a cumulative operator
*/
public void initializeCumulative(@NotNull final Context context, final long firstUnmodifiedKey,
final long firstUnmodifiedTimestamp,
@NotNull final RowSet bucketRowSet) {
context.reset();
}
/**
* Initialize the bucket context for a windowed operator
*/
public void initializeRolling(@NotNull final Context context,
@NotNull final RowSet bucketRowSet) {
context.reset();
}
/**
* Get the names of the input column(s) for this operator.
*
* @return the names of the input column
*/
@NotNull
protected String[] getInputColumnNames() {
return new String[] {pair.rightColumn};
}
/**
* Get the name of the timestamp column for this operator (or null if the operator does not require timestamp data).
*
* @return the name of the timestamp column
*/
@Nullable
protected String getTimestampColumnName() {
return timestampColumnName;
}
/**
* Get the value of the backward-looking window (might be nanoseconds or ticks).
*/
protected long getPrevWindowUnits() {
return reverseWindowScaleUnits;
}
/**
* Get the value of the forward-looking window (might be nanoseconds or ticks).
*/
protected long getFwdWindowUnits() {
return forwardWindowScaleUnits;
}
/**
* Get an array of column names that, when modified, affect the result of this computation.
*
* @return an array of column names that affect this operator.
*/
@NotNull
protected String[] getAffectingColumnNames() {
return affectingColumns;
}
/**
* Get an array of the output column names.
*
* @return the output column names.
*/
@NotNull
protected String[] getOutputColumnNames() {
return new String[] {pair.leftColumn};
}
/**
* Get a map of outputName to output {@link ColumnSource} for this operation.
*
* @return a map of output column name to output column source
*/
@NotNull
protected abstract Map> getOutputColumns();
/**
* Indicate that the operation should start tracking previous values for ticking updates.
*/
protected abstract void startTrackingPrev();
/**
* Make an {@link Context} suitable for use with updates.
*
* @param affectedChunkSize The maximum size of affected chunks that will be provided during the update.
* @param influencerChunkSize The maximum size of influencer chunks that will be provided during the update.
* @return a new context
*/
@NotNull
public abstract Context makeUpdateContext(final int affectedChunkSize, final int influencerChunkSize);
/**
* Perform any bookkeeping required at the end of a single part of the update. This is always preceded with a call
* to {@code #initializeUpdate(Context)} (specialized for each type of operator)
*
* @param context the context object
*/
@SuppressWarnings("unused")
protected void finishUpdate(@NotNull final Context context) {}
/**
* Apply a shift to the operation.
*/
protected abstract void applyOutputShift(@NotNull final RowSet subRowSetToShift, final long delta);
/**
* Prepare this operator output column for parallel updated.
*/
protected abstract void prepareForParallelPopulation(final RowSet changedRows);
/**
* Create the modified column set for the input columns of this operator.
*/
protected void createInputModifiedColumnSet(@NotNull final QueryTable source) {
inputModifiedColumnSet = source.newModifiedColumnSet(getAffectingColumnNames());
}
/**
* Create the modified column set for the output columns from this operator.
*/
protected void createOutputModifiedColumnSet(@NotNull final QueryTable result) {
outputModifiedColumnSet = result.newModifiedColumnSet(getOutputColumnNames());
}
/**
* Return the modified column set for the input columns of this operator.
*/
protected ModifiedColumnSet getInputModifiedColumnSet() {
return inputModifiedColumnSet;
}
/**
* Return the modified column set for the output columns from this operator.
*/
protected ModifiedColumnSet getOutputModifiedColumnSet() {
return outputModifiedColumnSet;
}
/**
* Set the downstream modified column set appropriately for this operator.
*/
protected void extractDownstreamModifiedColumnSet(@NotNull final TableUpdate upstream,
@NotNull final TableUpdate downstream) {
// for nearly all operators, all output columns will be modified.
downstream.modifiedColumnSet().setAll(getOutputModifiedColumnSet());
}
/**
* Clear the output rows by setting value to NULL. Dense sources will apply removes to the inner source.
*/
protected abstract void clearOutputRows(RowSet toClear);
/**
* Return whether the operator needs affected and influencer row positions during accumulation. Defaults to
* {@code false}.
*/
protected boolean requiresRowPositions() {
return false;
}
}
© 2015 - 2024 Weber Informatics LLC | Privacy Policy