Skip to content

armgit / solr

Public Imported from GitHub
0

Apache Solr open-source search software

SOLR-15044: JSON Loading: nested docs don't need ID (#143)

David Smiley5a9a34daedb3
diff --git a/solr/CHANGES.txt b/solr/CHANGES.txt
index 0e9a27b99b4..a90594b0ea1 100644
--- a/solr/CHANGES.txt
+++ b/solr/CHANGES.txt
@@ -411,6 +411,10 @@ Improvements
* SOLR-15156: [child] doc transformer's childFilter param no longer applies query syntax escaping.
It was inconsistent with other Solr options. It did this escaping since 8.0 but not prior. (David Smiley)
+* SOLR-15044: When indexing nested docs via JSON, it is no longer necessary to provide child doc IDs.
+ This was already working for XML & "javabin"/SolrJ. Previously, omitting the ID would be confused
+ for a partial/atomic update. (David Smiley)
+
Optimizations
---------------------
* SOLR-15079: Block Collapse - Faster collapse code when groups are co-located via Block Join style nested doc indexing.
diff --git a/solr/core/src/java/org/apache/solr/handler/loader/JsonLoader.java b/solr/core/src/java/org/apache/solr/handler/loader/JsonLoader.java
index 6791f0a84f8..cdb92b4c21b 100644
--- a/solr/core/src/java/org/apache/solr/handler/loader/JsonLoader.java
+++ b/solr/core/src/java/org/apache/solr/handler/loader/JsonLoader.java
@@ -22,7 +22,7 @@ import java.io.StringReader;
import java.lang.invoke.MethodHandles;
import java.util.ArrayList;
import java.util.Arrays;
-import java.util.HashMap;
+import java.util.Collections;
import java.util.Iterator;
import java.util.LinkedHashMap;
import java.util.List;
@@ -617,21 +617,26 @@ public class JsonLoader extends ContentStreamLoader {
assert ev == JSONParser.OBJECT_START;
SolrInputDocument extendedSolrDocument = parseDoc(ev);
- // is this a partial update or a child doc?
+ // Is this a child doc (true) or a partial update (false)
if (isChildDoc(extendedSolrDocument)) {
return extendedSolrDocument;
- } else {
- //return extendedSolrDocument.toMap(new HashMap<>(extendedSolrDocument.size())); not quite right
- Map<String, Object> map = new HashMap<>(extendedSolrDocument.size());
- for (SolrInputField inputField : extendedSolrDocument) {
- map.put(inputField.getName(), inputField.getValue());
- }
- return map;
+ } else { // partial update
+ assert extendedSolrDocument.size() == 1;
+ final SolrInputField pair = extendedSolrDocument.iterator().next();
+ return Collections.singletonMap(pair.getName(), pair.getValue());
}
}
+ /** Is this a child doc (true) or a partial update (false)? */
private boolean isChildDoc(SolrInputDocument extendedFieldValue) {
- return extendedFieldValue.containsKey(req.getSchema().getUniqueKeyField().getName());
+ if (extendedFieldValue.size() != 1) {
+ return true;
+ }
+ // if the only key is a field in the schema, assume it's a child doc
+ final String fieldName = extendedFieldValue.iterator().next().getName();
+ return req.getSchema().getFieldOrNull(fieldName) != null;
+ // otherwise, assume it's "set" or some other verb for a partial update.
+ // NOTE: it's fundamentally ambiguous with JSON; this is a best effort try.
}
}
diff --git a/solr/core/src/test/org/apache/solr/handler/JsonLoaderTest.java b/solr/core/src/test/org/apache/solr/handler/JsonLoaderTest.java
index cc917765f15..1895ba0445a 100644
--- a/solr/core/src/test/org/apache/solr/handler/JsonLoaderTest.java
+++ b/solr/core/src/test/org/apache/solr/handler/JsonLoaderTest.java
@@ -791,6 +791,14 @@ public class JsonLoaderTest extends SolrTestCaseJ4 {
checkTwoAnonymousChildDocs(DUP_KEYS_ANON_CHILD_DOCS_JSON, false);
}
+ @Test
+ public void testChildDocWithoutId() throws Exception {
+ final String json = DUP_KEYS_ANON_CHILD_DOCS_JSON.replace("\"id\": \"3\",\n", "");
+ assert !json.equals(DUP_KEYS_ANON_CHILD_DOCS_JSON);
+ checkTwoAnonymousChildDocs(
+ json, false);
+ }
+
// rawJsonStr has "_childDocuments_" key. if anonChildDocs then we want to test with something else.
private void checkTwoAnonymousChildDocs(String rawJsonStr, boolean anonChildDocs) throws Exception {
if (!anonChildDocs) {
@@ -824,7 +832,11 @@ public class JsonLoaderTest extends SolrTestCaseJ4 {
cd = (SolrInputDocument)((List)(d.getField("childLabel")).getValue()).get(1);
}
cf = cd.getField( "id" );
- assertEquals("3", cf.getValue());
+ if (rawJsonStr.contains("\"3\"")) {
+ assertEquals("3", cf.getValue());
+ } else { // ID 3 was removed previously to test we don't need an ID to have a child doc
+ assertNull("child doc should have no ID", cf);
+ }
cf = cd.getField( "foo_i" );
assertEquals(2, cf.getValueCount());