Skip to content

Respect include/exclude patterns when applying exclude_vectors #135444

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 1 commit into from
Sep 25, 2025
Merged

Respect include/exclude patterns when applying exclude_vectors #135444

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
1 change: 0 additions & 1 deletion ...stTest/resources/rest-api-spec/test/search.vectors/240_source_synthetic_dense_vectors.yml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -153,7 +153,6 @@ setup:
index: test
body:
_source:
exclude_vectors: false
includes: nested.vector
sort: ["name"]

Expand Down
57 changes: 56 additions & 1 deletion ...tTest/resources/rest-api-spec/test/search.vectors/250_source_synthetic_sparse_vectors.yml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -165,15 +165,49 @@ setup:
index: test
body:
_source:
exclude_vectors: false
includes: nested.emb
sort: ["name"]

- match: { hits.hits.0._id: "1"}
- length: { hits.hits.0._source: 0}

- match: { hits.hits.1._id: "2"}
- length: { hits.hits.1._source: 1 }
- length: { hits.hits.1._source.nested: 3 }
- exists: hits.hits.1._source.nested.0.emb
- not_exists: hits.hits.1._source.nested.0.paragraph_id
- exists: hits.hits.1._source.nested.1.emb
- not_exists: hits.hits.1._source.nested.1.paragraph_id
- exists: hits.hits.1._source.nested.2.emb
- not_exists: hits.hits.1._source.nested.2.paragraph_id

- match: { hits.hits.2._id: "3" }
- length: { hits.hits.2._source: 0}

- match: { hits.hits.3._id: "4" }
- length: { hits.hits.3._source: 1 }
- length: { hits.hits.3._source.nested: 2 }
- exists: hits.hits.3._source.nested.0.emb
- length: { hits.hits.3._source.nested.0.emb: 3 }
- not_exists: hits.hits.3._source.nested.0.paragraph_id
- exists: hits.hits.3._source.nested.1.emb
- length: { hits.hits.3._source.nested.1.emb: 1 }
- not_exists: hits.hits.3._source.nested.1.paragraph_id

- do:
search:
index: test
body:
_source:
exclude_vectors: true
includes: nested.emb
sort: ["name"]

- match: { hits.hits.0._id: "1"}
- length: { hits.hits.0._source: 0}

- match: { hits.hits.1._id: "2"}
- length: { hits.hits.1._source: 1 }
- length: { hits.hits.1._source.nested: 3 }
- exists: hits.hits.1._source.nested.0.emb
- not_exists: hits.hits.1._source.nested.0.paragraph_id
Expand All @@ -195,6 +229,27 @@ setup:
- length: { hits.hits.3._source.nested.1.emb: 1 }
- not_exists: hits.hits.3._source.nested.1.paragraph_id

- do:
search:
index: test
body:
_source:
exclude_vectors: true
includes: nested.emb
excludes: nested*
sort: ["name"]

- match: { hits.hits.0._id: "1"}
- length: { hits.hits.0._source: 0}

- match: { hits.hits.1._id: "2"}
- length: { hits.hits.1._source: 0 }
- match: { hits.hits.2._id: "3" }
- length: { hits.hits.2._source: 0}

- match: { hits.hits.3._id: "4" }
- length: { hits.hits.3._source: 0 }

- do:
headers:
# Force JSON content type so that we use a parser that interprets the embeddings as doubles
Expand Down
6 changes: 6 additions & 0 deletions server/src/main/java/org/elasticsearch/index/get/ShardGetService.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -458,8 +458,14 @@ public static Tuple<FetchSourceContext, SourceFilter> maybeExcludeVectorFields(
)
: null;

SourceFilter filter = fetchSourceContext != null ? fetchSourceContext.filter() : null;

List<String> lateExcludes = new ArrayList<>();
var excludes = mappingLookup.getFullNameToFieldType().values().stream().filter(MappedFieldType::isVectorEmbedding).filter(f -> {
// Keep the vector fields that are explicitly included and not explicitly excluded
if (filter != null && filter.isExplicitlyIncluded(f.name())) {
return filter.isPathFiltered(f.name(), false);
}
// Exclude the field specified by the `fields` option
if (fetchFieldsAut != null && fetchFieldsAut.run(f.name())) {
lateExcludes.add(f.name());
Expand Down
24 changes: 22 additions & 2 deletions server/src/main/java/org/elasticsearch/search/lookup/SourceFilter.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -68,6 +68,26 @@ public String[] getExcludes() {
return excludes;
}

/**
* Checks if the given path matches at least one explicitly defined include pattern.
* <p>
* If no include patterns are defined, this method always returns {@code false}.
*
* @param fullPath the full path to evaluate
* @return {@code true} if the path matches any explicitly defined include pattern,
* {@code false} otherwise
*/
public boolean isExplicitlyIncluded(String fullPath) {
if (includes.length == 0) {
return false;
}
if (includeAut == null) {
includeAut = XContentMapValues.compileAutomaton(includes, new CharacterRunAutomaton(Automata.makeAnyString()));
}
int state = step(includeAut, fullPath, 0);
return state != -1 && includeAut.isAccept(state);
}

/**
* Determines whether the given full path should be filtered out.
*
Expand All @@ -77,7 +97,7 @@ public String[] getExcludes() {
*/
public boolean isPathFiltered(String fullPath, boolean isObject) {
final boolean included;
if (includes != null) {
if (includes.length > 0) {
if (includeAut == null) {
includeAut = XContentMapValues.compileAutomaton(includes, new CharacterRunAutomaton(Automata.makeAnyString()));
}
Expand All @@ -87,7 +107,7 @@ public boolean isPathFiltered(String fullPath, boolean isObject) {
included = true;
}

if (excludes != null) {
if (excludes.length > 0) {
if (excludeAut == null) {
excludeAut = XContentMapValues.compileAutomaton(excludes, new CharacterRunAutomaton(Automata.makeEmpty()));
}
Expand Down
50 changes: 50 additions & 0 deletions server/src/test/java/org/elasticsearch/search/lookup/SourceFilterTests.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -156,4 +156,54 @@ public void testIncludeParentAndExcludeChildSubFieldsObjects() {
Source filteredBytes = fromBytes.filter(new SourceFilter(new String[] { "myObject" }, new String[] { "myObject.myField" }));
assertEquals(filteredBytes.source(), Map.of("myObject", Map.of("other", "otherValue")));
}

public void testIsExplicitlyIncluded() {
var filter = new SourceFilter(null, null);
assertFalse(filter.isExplicitlyIncluded("foo"));

filter = new SourceFilter(new String[] {}, null);
assertFalse(filter.isExplicitlyIncluded("foo"));

filter = new SourceFilter(new String[] { "foo", "bar.*" }, null);
assertTrue(filter.isExplicitlyIncluded("foo"));
assertTrue(filter.isExplicitlyIncluded("bar.field"));
assertFalse(filter.isExplicitlyIncluded("baz"));
assertFalse(filter.isExplicitlyIncluded("bar"));
}

public void testIsPathFilteredWithExcludes() {
var filter = new SourceFilter(null, new String[] { "foo", "bar.field" });
assertTrue(filter.isPathFiltered("foo", true));
assertTrue(filter.isPathFiltered("foo", false));

assertTrue(filter.isPathFiltered("bar.field", false));
assertFalse(filter.isPathFiltered("baz", false));
assertFalse(filter.isPathFiltered("bar", false));
assertFalse(filter.isPathFiltered("bar", true));
}

public void testIsPathFilteredWithIncludes() {
var filter = new SourceFilter(new String[] { "foo", "bar.field" }, null);
assertFalse(filter.isPathFiltered("foo", true));
assertFalse(filter.isPathFiltered("foo", false));

assertFalse(filter.isPathFiltered("bar.field", false));
assertTrue(filter.isPathFiltered("baz", false));
assertTrue(filter.isPathFiltered("bar", false));
assertFalse(filter.isPathFiltered("bar", true));
}

public void testIsPathFilteredWithIncludesAndExcludes() {
var filter = new SourceFilter(new String[] { "foo", "bar.*", "nested.field" }, new String[] { "foo", "bar.field" });
assertTrue(filter.isPathFiltered("foo", true));
assertTrue(filter.isPathFiltered("foo", false));

assertTrue(filter.isPathFiltered("bar.field", false));
assertTrue(filter.isPathFiltered("baz", false));
assertTrue(filter.isPathFiltered("bar", false));
assertFalse(filter.isPathFiltered("bar", true));

assertFalse(filter.isPathFiltered("nested.field", false));
assertTrue(filter.isPathFiltered("nested.another", false));
}
}
1 change: 0 additions & 1 deletion ...yamlRestTest/resources/rest-api-spec/test/rank_vectors/rank_vectors_synthetic_vectors.yml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -154,7 +154,6 @@ setup:
index: test
body:
_source:
exclude_vectors: false
includes: nested.vector
sort: ["name"]

Expand Down