org.elasticsearch.index.get.ShardGetService Maven / Gradle / Ivy
Show all versions of elasticsearch Show documentation
/*
* Copyright Elasticsearch B.V. and/or licensed to Elasticsearch B.V. under one
* or more contributor license agreements. Licensed under the Elastic License
* 2.0 and the Server Side Public License, v 1; you may not use this file except
* in compliance with, at your election, the Elastic License 2.0 or the Server
* Side Public License, v 1.
*/
package org.elasticsearch.index.get;
import org.apache.lucene.index.DocValuesType;
import org.apache.lucene.index.FieldInfo;
import org.apache.lucene.index.IndexOptions;
import org.apache.lucene.index.IndexableField;
import org.apache.lucene.index.IndexableFieldType;
import org.apache.lucene.index.StoredFieldVisitor;
import org.apache.lucene.index.Term;
import org.elasticsearch.ElasticsearchException;
import org.elasticsearch.common.Nullable;
import org.elasticsearch.common.bytes.BytesReference;
import org.elasticsearch.common.collect.Tuple;
import org.elasticsearch.common.document.DocumentField;
import org.elasticsearch.common.lucene.uid.Versions;
import org.elasticsearch.common.lucene.uid.VersionsAndSeqNoResolver.DocIdAndVersion;
import org.elasticsearch.common.metrics.CounterMetric;
import org.elasticsearch.common.metrics.MeanMetric;
import org.elasticsearch.common.util.set.Sets;
import org.elasticsearch.common.xcontent.XContentFactory;
import org.elasticsearch.common.xcontent.XContentHelper;
import org.elasticsearch.common.xcontent.XContentType;
import org.elasticsearch.common.xcontent.support.XContentMapValues;
import org.elasticsearch.index.IndexSettings;
import org.elasticsearch.index.VersionType;
import org.elasticsearch.index.engine.Engine;
import org.elasticsearch.index.engine.TranslogLeafReader;
import org.elasticsearch.index.fieldvisitor.CustomFieldsVisitor;
import org.elasticsearch.index.fieldvisitor.FieldsVisitor;
import org.elasticsearch.index.mapper.DocumentMapper;
import org.elasticsearch.index.mapper.IdFieldMapper;
import org.elasticsearch.index.mapper.Mapper;
import org.elasticsearch.index.mapper.MapperService;
import org.elasticsearch.index.mapper.ParsedDocument;
import org.elasticsearch.index.mapper.RoutingFieldMapper;
import org.elasticsearch.index.mapper.SourceFieldMapper;
import org.elasticsearch.index.mapper.SourceToParse;
import org.elasticsearch.index.mapper.Uid;
import org.elasticsearch.index.shard.AbstractIndexShardComponent;
import org.elasticsearch.index.shard.IndexShard;
import org.elasticsearch.search.fetch.subphase.FetchSourceContext;
import java.io.IOException;
import java.util.Collections;
import java.util.HashMap;
import java.util.List;
import java.util.Map;
import java.util.concurrent.TimeUnit;
import java.util.stream.Stream;
import static org.elasticsearch.index.seqno.SequenceNumbers.UNASSIGNED_PRIMARY_TERM;
import static org.elasticsearch.index.seqno.SequenceNumbers.UNASSIGNED_SEQ_NO;
public final class ShardGetService extends AbstractIndexShardComponent {
private final MapperService mapperService;
private final MeanMetric existsMetric = new MeanMetric();
private final MeanMetric missingMetric = new MeanMetric();
private final CounterMetric currentMetric = new CounterMetric();
private final IndexShard indexShard;
public ShardGetService(IndexSettings indexSettings, IndexShard indexShard,
MapperService mapperService) {
super(indexShard.shardId(), indexSettings);
this.mapperService = mapperService;
this.indexShard = indexShard;
}
public GetStats stats() {
return new GetStats(existsMetric.count(), TimeUnit.NANOSECONDS.toMillis(existsMetric.sum()),
missingMetric.count(), TimeUnit.NANOSECONDS.toMillis(missingMetric.sum()), currentMetric.count());
}
public GetResult get(String type, String id, String[] gFields, boolean realtime, long version,
VersionType versionType, FetchSourceContext fetchSourceContext) {
return
get(type, id, gFields, realtime, version, versionType, UNASSIGNED_SEQ_NO, UNASSIGNED_PRIMARY_TERM, fetchSourceContext);
}
private GetResult get(String type, String id, String[] gFields, boolean realtime, long version, VersionType versionType,
long ifSeqNo, long ifPrimaryTerm, FetchSourceContext fetchSourceContext) {
currentMetric.inc();
try {
long now = System.nanoTime();
GetResult getResult =
innerGet(type, id, gFields, realtime, version, versionType, ifSeqNo, ifPrimaryTerm, fetchSourceContext);
if (getResult.isExists()) {
existsMetric.inc(System.nanoTime() - now);
} else {
missingMetric.inc(System.nanoTime() - now);
}
return getResult;
} finally {
currentMetric.dec();
}
}
public GetResult getForUpdate(String type, String id, long ifSeqNo, long ifPrimaryTerm) {
return get(type, id, new String[]{RoutingFieldMapper.NAME}, true,
Versions.MATCH_ANY, VersionType.INTERNAL, ifSeqNo, ifPrimaryTerm, FetchSourceContext.FETCH_SOURCE);
}
/**
* Returns {@link GetResult} based on the specified {@link org.elasticsearch.index.engine.Engine.GetResult} argument.
* This method basically loads specified fields for the associated document in the engineGetResult.
* This method load the fields from the Lucene index and not from transaction log and therefore isn't realtime.
*
* Note: Call must release engine searcher associated with engineGetResult!
*/
public GetResult get(Engine.GetResult engineGetResult, String id, String type,
String[] fields, FetchSourceContext fetchSourceContext) {
if (engineGetResult.exists() == false) {
return new GetResult(shardId.getIndexName(), type, id, UNASSIGNED_SEQ_NO, UNASSIGNED_PRIMARY_TERM, -1, false, null, null, null);
}
currentMetric.inc();
try {
long now = System.nanoTime();
fetchSourceContext = normalizeFetchSourceContent(fetchSourceContext, fields);
GetResult getResult = innerGetLoadFromStoredFields(type, id, fields, fetchSourceContext, engineGetResult, mapperService);
if (getResult.isExists()) {
existsMetric.inc(System.nanoTime() - now);
} else {
missingMetric.inc(System.nanoTime() - now); // This shouldn't happen...
}
return getResult;
} finally {
currentMetric.dec();
}
}
/**
* decides what needs to be done based on the request input and always returns a valid non-null FetchSourceContext
*/
private FetchSourceContext normalizeFetchSourceContent(@Nullable FetchSourceContext context, @Nullable String[] gFields) {
if (context != null) {
return context;
}
if (gFields == null) {
return FetchSourceContext.FETCH_SOURCE;
}
for (String field : gFields) {
if (SourceFieldMapper.NAME.equals(field)) {
return FetchSourceContext.FETCH_SOURCE;
}
}
return FetchSourceContext.DO_NOT_FETCH_SOURCE;
}
private GetResult innerGet(String type, String id, String[] gFields, boolean realtime, long version, VersionType versionType,
long ifSeqNo, long ifPrimaryTerm, FetchSourceContext fetchSourceContext) {
fetchSourceContext = normalizeFetchSourceContent(fetchSourceContext, gFields);
if (type == null || type.equals("_all")) {
DocumentMapper mapper = mapperService.documentMapper();
type = mapper == null ? null : mapper.type();
}
Engine.GetResult get = null;
if (type != null) {
Term uidTerm = new Term(IdFieldMapper.NAME, Uid.encodeId(id));
get = indexShard.get(new Engine.Get(realtime, realtime, type, id, uidTerm)
.version(version).versionType(versionType).setIfSeqNo(ifSeqNo).setIfPrimaryTerm(ifPrimaryTerm));
assert get.isFromTranslog() == false || realtime : "should only read from translog if realtime enabled";
if (get.exists() == false) {
get.close();
}
}
if (get == null || get.exists() == false) {
return new GetResult(shardId.getIndexName(), type, id, UNASSIGNED_SEQ_NO, UNASSIGNED_PRIMARY_TERM, -1, false, null, null, null);
}
try {
// break between having loaded it from translog (so we only have _source), and having a document to load
return innerGetLoadFromStoredFields(type, id, gFields, fetchSourceContext, get, mapperService);
} finally {
get.close();
}
}
private GetResult innerGetLoadFromStoredFields(String type, String id, String[] storedFields, FetchSourceContext fetchSourceContext,
Engine.GetResult get, MapperService mapperService) {
assert get.exists() : "method should only be called if document could be retrieved";
// check first if stored fields to be loaded don't contain an object field
DocumentMapper docMapper = mapperService.documentMapper();
if (storedFields != null) {
for (String field : storedFields) {
Mapper fieldMapper = docMapper.mappers().getMapper(field);
if (fieldMapper == null) {
if (docMapper.mappers().objectMappers().get(field) != null) {
// Only fail if we know it is a object field, missing paths / fields shouldn't fail.
throw new IllegalArgumentException("field [" + field + "] isn't a leaf field");
}
}
}
}
Map documentFields = null;
Map metadataFields = null;
BytesReference source = null;
DocIdAndVersion docIdAndVersion = get.docIdAndVersion();
// force fetching source if we read from translog and need to recreate stored fields
boolean forceSourceForComputingTranslogStoredFields = get.isFromTranslog() && storedFields != null &&
Stream.of(storedFields).anyMatch(f -> TranslogLeafReader.ALL_FIELD_NAMES.contains(f) == false);
FieldsVisitor fieldVisitor = buildFieldsVisitors(storedFields,
forceSourceForComputingTranslogStoredFields ? FetchSourceContext.FETCH_SOURCE : fetchSourceContext);
if (fieldVisitor != null) {
try {
docIdAndVersion.reader.document(docIdAndVersion.docId, fieldVisitor);
} catch (IOException e) {
throw new ElasticsearchException("Failed to get type [" + type + "] and id [" + id + "]", e);
}
source = fieldVisitor.source();
// in case we read from translog, some extra steps are needed to make _source consistent and to load stored fields
if (get.isFromTranslog()) {
// Fast path: if only asked for the source or stored fields that have been already provided by TranslogLeafReader,
// just make source consistent by reapplying source filters from mapping (possibly also nulling the source)
if (forceSourceForComputingTranslogStoredFields == false) {
try {
source = indexShard.mapperService().documentMapper().sourceMapper().applyFilters(source, null);
} catch (IOException e) {
throw new ElasticsearchException("Failed to reapply filters for [" + id + "] after reading from translog", e);
}
} else {
// Slow path: recreate stored fields from original source
assert source != null : "original source in translog must exist";
SourceToParse sourceToParse = new SourceToParse(shardId.getIndexName(), type, id, source,
XContentHelper.xContentType(source), fieldVisitor.routing(), Collections.emptyMap());
ParsedDocument doc = indexShard.mapperService().documentMapper().parse(sourceToParse);
assert doc.dynamicMappingsUpdate() == null : "mapping updates should not be required on already-indexed doc";
// update special fields
doc.updateSeqID(docIdAndVersion.seqNo, docIdAndVersion.primaryTerm);
doc.version().setLongValue(docIdAndVersion.version);
// retrieve stored fields from parsed doc
fieldVisitor = buildFieldsVisitors(storedFields, fetchSourceContext);
for (IndexableField indexableField : doc.rootDoc().getFields()) {
IndexableFieldType fieldType = indexableField.fieldType();
if (fieldType.stored()) {
FieldInfo fieldInfo = new FieldInfo(indexableField.name(), 0, false, false, false, IndexOptions.NONE,
DocValuesType.NONE, -1, Collections.emptyMap(), 0, 0, 0, false);
StoredFieldVisitor.Status status = fieldVisitor.needsField(fieldInfo);
if (status == StoredFieldVisitor.Status.YES) {
if (indexableField.numericValue() != null) {
fieldVisitor.objectField(fieldInfo, indexableField.numericValue());
} else if (indexableField.binaryValue() != null) {
fieldVisitor.binaryField(fieldInfo, indexableField.binaryValue());
} else if (indexableField.stringValue() != null) {
fieldVisitor.objectField(fieldInfo, indexableField.stringValue());
}
} else if (status == StoredFieldVisitor.Status.STOP) {
break;
}
}
}
// retrieve source (with possible transformations, e.g. source filters
source = fieldVisitor.source();
}
}
// put stored fields into result objects
if (fieldVisitor.fields().isEmpty() == false) {
fieldVisitor.postProcess(mapperService::fieldType,
mapperService.documentMapper() == null ? null : mapperService.documentMapper().type());
documentFields = new HashMap<>();
metadataFields = new HashMap<>();
for (Map.Entry> entry : fieldVisitor.fields().entrySet()) {
if (mapperService.isMetadataField(entry.getKey())) {
metadataFields.put(entry.getKey(), new DocumentField(entry.getKey(), entry.getValue()));
} else {
documentFields.put(entry.getKey(), new DocumentField(entry.getKey(), entry.getValue()));
}
}
}
}
if (source != null) {
// apply request-level source filtering
if (fetchSourceContext.fetchSource() == false) {
source = null;
} else if (fetchSourceContext.includes().length > 0 || fetchSourceContext.excludes().length > 0) {
Map sourceAsMap;
// TODO: The source might be parsed and available in the sourceLookup but that one uses unordered maps so different.
// Do we care?
Tuple> typeMapTuple = XContentHelper.convertToMap(source, true);
XContentType sourceContentType = typeMapTuple.v1();
sourceAsMap = typeMapTuple.v2();
sourceAsMap = XContentMapValues.filter(sourceAsMap, fetchSourceContext.includes(), fetchSourceContext.excludes());
try {
source = BytesReference.bytes(XContentFactory.contentBuilder(sourceContentType).map(sourceAsMap));
} catch (IOException e) {
throw new ElasticsearchException("Failed to get id [" + id + "] with includes/excludes set", e);
}
}
}
if (fetchSourceContext.fetchSource() == false) {
source = null;
}
if (source != null && get.isFromTranslog()) {
// reapply source filters from mapping (possibly also nulling the source)
try {
source = docMapper.sourceMapper().applyFilters(source, null);
} catch (IOException e) {
throw new ElasticsearchException("Failed to reapply filters for [" + id + "] after reading from translog", e);
}
}
if (source != null && (fetchSourceContext.includes().length > 0 || fetchSourceContext.excludes().length > 0)) {
Map sourceAsMap;
// TODO: The source might parsed and available in the sourceLookup but that one uses unordered maps so different. Do we care?
Tuple> typeMapTuple = XContentHelper.convertToMap(source, true);
XContentType sourceContentType = typeMapTuple.v1();
sourceAsMap = typeMapTuple.v2();
sourceAsMap = XContentMapValues.filter(sourceAsMap, fetchSourceContext.includes(), fetchSourceContext.excludes());
try {
source = BytesReference.bytes(XContentFactory.contentBuilder(sourceContentType).map(sourceAsMap));
} catch (IOException e) {
throw new ElasticsearchException("Failed to get type [" + type + "] and id [" + id + "] with includes/excludes set", e);
}
}
return new GetResult(shardId.getIndexName(), type, id, get.docIdAndVersion().seqNo, get.docIdAndVersion().primaryTerm,
get.version(), get.exists(), source, documentFields, metadataFields);
}
private static FieldsVisitor buildFieldsVisitors(String[] fields, FetchSourceContext fetchSourceContext) {
if (fields == null || fields.length == 0) {
return fetchSourceContext.fetchSource() ? new FieldsVisitor(true) : null;
}
return new CustomFieldsVisitor(Sets.newHashSet(fields), fetchSourceContext.fetchSource());
}
}