io.deephaven.engine.table.impl.updateby.UpdateByWindowCumulative 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.base.verify.Assert;
import io.deephaven.chunk.Chunk;
import io.deephaven.chunk.LongChunk;
import io.deephaven.chunk.attributes.Values;
import io.deephaven.engine.rowset.*;
import io.deephaven.engine.table.ChunkSource;
import io.deephaven.engine.table.ColumnSource;
import io.deephaven.engine.table.TableUpdate;
import io.deephaven.engine.table.impl.ssa.LongSegmentedSortedArray;
import org.jetbrains.annotations.NotNull;
import org.jetbrains.annotations.Nullable;
import java.util.BitSet;
import java.util.stream.IntStream;
import static io.deephaven.engine.rowset.RowSequence.NULL_ROW_KEY;
import static io.deephaven.util.QueryConstants.NULL_LONG;
/**
* This is the specialization of {@link UpdateByWindow} that handles `cumulative` operators. These operators do not
* maintain a window of data and can be computed from the previous value and the current value.
*/
class UpdateByWindowCumulative extends UpdateByWindow {
UpdateByWindowCumulative(UpdateByOperator[] operators, int[][] operatorSourceSlots,
@Nullable String timestampColumnName) {
super(operators, operatorSourceSlots, timestampColumnName);
}
@Override
void prepareWindowBucket(UpdateByWindowBucketContext context) {
// working chunk size need not be larger than affectedRows.size()
context.workingChunkSize = Math.toIntExact(Math.min(context.workingChunkSize, context.affectedRows.size()));
}
@Override
void finalizeWindowBucket(UpdateByWindowBucketContext context) {
super.finalizeWindowBucket(context);
}
@Override
UpdateByWindowBucketContext makeWindowContext(final TrackingRowSet sourceRowSet,
final ColumnSource timestampColumnSource,
final LongSegmentedSortedArray timestampSsa,
final TrackingRowSet timestampValidRowSet,
final boolean timestampsModified,
final int chunkSize,
final boolean isInitializeStep) {
return new UpdateByWindowBucketContext(sourceRowSet, timestampColumnSource, timestampSsa, timestampValidRowSet,
timestampsModified, chunkSize, isInitializeStep);
}
@Override
void computeAffectedRowsAndOperators(UpdateByWindowBucketContext context, @NotNull TableUpdate upstream) {
if (upstream.empty() || context.sourceRowSet.isEmpty()) {
// No further work will be done on this context
finalizeWindowBucket(context);
return;
}
// all rows are affected on the initial step
if (context.initialStep) {
context.affectedRows = context.sourceRowSet.copy();
context.influencerRows = context.affectedRows;
// mark all operators as affected by this update
context.dirtyOperatorIndices = IntStream.range(0, operators.length).toArray();
context.dirtyOperators = new BitSet(operators.length);
context.dirtyOperators.set(0, operators.length);
context.isDirty = true;
return;
}
// determine which operators are affected by this update
processUpdateForContext(context, upstream);
if (!context.isDirty) {
// No further work will be done on this context
finalizeWindowBucket(context);
return;
}
// we can ignore modifications if they do not affect our input columns
final boolean inputModified = context.inputModified
|| (timestampColumnName != null && context.timestampsModified);
long smallestModifiedKey = smallestAffectedKey(upstream, context.sourceRowSet, inputModified);
context.affectedRows = smallestModifiedKey == Long.MAX_VALUE
? RowSetFactory.empty()
: context.sourceRowSet.subSetByKeyRange(smallestModifiedKey, context.sourceRowSet.lastRowKey());
if (context.affectedRows.isEmpty()) {
// No further work will be done on this context
finalizeWindowBucket(context);
context.isDirty = false;
return;
}
context.influencerRows = context.affectedRows;
}
@Override
void processWindowBucketOperatorSet(final UpdateByWindowBucketContext context,
final int[] opIndices,
final int[] srcIndices,
final UpdateByOperator.Context[] winOpContexts,
final Chunk[] chunkArr,
final ChunkSource.GetContext[] chunkContexts,
final boolean initialStep) {
Assert.neqNull(context.inputSources, "assignInputSources() must be called before processRow()");
// Identify an operator to use for determining initialization state.
final UpdateByOperator firstOp;
final UpdateByOperator.Context firstOpCtx;
if (initialStep || timestampColumnName == null) {
// We can use any operator. When initialStep==true, we start from the beginning.
firstOp = operators[opIndices[0]];
firstOpCtx = winOpContexts[0];
} else {
// Use the first time-sensitive operator if one exists, or first overall otherwise.
final int match = IntStream.range(0, opIndices.length)
.filter(ii -> operators[opIndices[ii]].timestampColumnName != null)
.findAny()
.orElse(0);
firstOp = operators[opIndices[match]];
firstOpCtx = winOpContexts[match];
}
try (final RowSequence.Iterator affectedIt = context.affectedRows.getRowSequenceIterator();
ChunkSource.GetContext tsGetContext =
context.timestampColumnSource == null ? null
: context.timestampColumnSource.makeGetContext(context.workingChunkSize)) {
final long rowKey;
final long timestamp;
if (initialStep) {
// We are always at the beginning of the RowSet at creation phase.
rowKey = NULL_ROW_KEY;
timestamp = NULL_LONG;
} else {
// Find the key before the first affected row.
final long pos = context.sourceRowSet.find(context.affectedRows.firstRowKey());
final long keyBefore = pos == 0 ? NULL_ROW_KEY : context.sourceRowSet.get(pos - 1);
// Preload that data for these operators.
if (firstOp.timestampColumnName == null || keyBefore == NULL_ROW_KEY) {
// This operator doesn't care about timestamps or we know we are at the beginning of the rowset
rowKey = keyBefore;
timestamp = NULL_LONG;
} else {
// This operator cares about timestamps, so make sure it is starting from a valid value and
// valid timestamp by looking backward until the conditions are met.
long potentialResetTimestamp = context.timestampColumnSource.getLong(keyBefore);
if (potentialResetTimestamp == NULL_LONG || !firstOpCtx.isValueValid(keyBefore)) {
try (final RowSet.SearchIterator rIt = context.sourceRowSet.reverseIterator()) {
if (rIt.advance(keyBefore)) {
while (rIt.hasNext()) {
final long nextKey = rIt.nextLong();
potentialResetTimestamp = context.timestampColumnSource.getLong(nextKey);
if (potentialResetTimestamp != NULL_LONG &&
firstOpCtx.isValueValid(nextKey)) {
break;
}
}
}
}
}
rowKey = keyBefore;
timestamp = potentialResetTimestamp;
}
}
// Call the specialized version of `intializeUpdate()` for these operators.
for (int ii = 0; ii < opIndices.length; ii++) {
final int opIdx = opIndices[ii];
if (!context.dirtyOperators.get(opIdx)) {
// Skip if not dirty.
continue;
}
UpdateByOperator cumOp = operators[opIdx];
cumOp.initializeCumulative(winOpContexts[ii], rowKey, timestamp, context.sourceRowSet);
}
while (affectedIt.hasMore()) {
final RowSequence affectedRs = affectedIt.getNextRowSequenceWithLength(context.workingChunkSize);
// Create the timestamp chunk if needed.
LongChunk tsChunk = context.timestampColumnSource == null ? null
: context.timestampColumnSource.getChunk(tsGetContext, affectedRs).asLongChunk();
// Prep the chunk array needed by the accumulate call.
for (int ii = 0; ii < srcIndices.length; ii++) {
int srcIdx = srcIndices[ii];
chunkArr[ii] = context.inputSources[srcIdx].getChunk(chunkContexts[ii], affectedRs);
}
// Make the specialized call for windowed operators.
for (int ii = 0; ii < opIndices.length; ii++) {
final int opIdx = opIndices[ii];
if (!context.dirtyOperators.get(opIdx)) {
// Skip if not dirty.
continue;
}
winOpContexts[ii].accumulateCumulative(
affectedRs,
chunkArr,
tsChunk,
affectedRs.intSize());
}
}
// Finalize the operator.
for (int ii = 0; ii < opIndices.length; ii++) {
final int opIdx = opIndices[ii];
if (!context.dirtyOperators.get(opIdx)) {
// Skip if not dirty.
continue;
}
UpdateByOperator cumOp = operators[opIdx];
cumOp.finishUpdate(winOpContexts[ii]);
}
}
}
/**
* Find the smallest valued key that participated in the upstream {@link TableUpdate}.
*
* @param upstream the {@link TableUpdate update} from upstream
* @param affectedRowSet the {@link TrackingRowSet rowset} for the current bucket
* @param inputModified whether the input columns for this window were modified
*
* @return the smallest key that participated in any part of the update. This will be the minimum of the first key
* of each of added, modified and removed (post-shift) rows.
*/
private static long smallestAffectedKey(
@NotNull final TableUpdate upstream,
@NotNull final TrackingRowSet affectedRowSet,
final boolean inputModified) {
long smallestModifiedKey = Long.MAX_VALUE;
if (upstream.removed().isNonempty()) {
// removed rows aren't represented in the shift data, so choose the row immediately preceding the first
// removed as the removed candidate for smallestAffectedKey
final long pos = affectedRowSet.findPrev(upstream.removed().firstRowKey());
if (pos == 0) {
// the first row was removed, recompute everything
return affectedRowSet.firstRowKey();
}
// get the key previous to this one and shift to post-space
smallestModifiedKey = affectedRowSet.getPrev(pos - 1);
if (upstream.shifted().nonempty()) {
smallestModifiedKey = upstream.shifted().apply(smallestModifiedKey);
}
// tighten this up more by advancing one key in the post-shift space. This leaves us with first key
// following the first remove
if (smallestModifiedKey < affectedRowSet.lastRowKey()) {
smallestModifiedKey = affectedRowSet.get(affectedRowSet.find(smallestModifiedKey) + 1);
} else {
// all removes are after the end of the current rowset
smallestModifiedKey = Long.MAX_VALUE;
}
}
if (upstream.added().isNonempty()) {
smallestModifiedKey = Math.min(smallestModifiedKey, upstream.added().firstRowKey());
}
// consider the modifications only when input columns were modified
if (upstream.modified().isNonempty() && inputModified) {
smallestModifiedKey = Math.min(smallestModifiedKey, upstream.modified().firstRowKey());
}
return smallestModifiedKey;
}
}
© 2015 - 2024 Weber Informatics LLC | Privacy Policy