You are viewing a plain text version of this content. The canonical link for it is here.
Posted to solr-commits@lucene.apache.org by yo...@apache.org on 2007/06/27 07:52:03 UTC

svn commit: r551060 - /lucene/solr/trunk/src/java/org/apache/solr/update/DocumentBuilder.java

Author: yonik
Date: Tue Jun 26 22:52:02 2007
New Revision: 551060

URL: http://svn.apache.org/viewvc?view=rev&rev=551060
Log:
check for null copyField source, delay some allocations

Modified:
    lucene/solr/trunk/src/java/org/apache/solr/update/DocumentBuilder.java

Modified: lucene/solr/trunk/src/java/org/apache/solr/update/DocumentBuilder.java
URL: http://svn.apache.org/viewvc/lucene/solr/trunk/src/java/org/apache/solr/update/DocumentBuilder.java?view=diff&rev=551060&r1=551059&r2=551060
==============================================================================
--- lucene/solr/trunk/src/java/org/apache/solr/update/DocumentBuilder.java (original)
+++ lucene/solr/trunk/src/java/org/apache/solr/update/DocumentBuilder.java Tue Jun 26 22:52:02 2007
@@ -43,7 +43,7 @@
 public class DocumentBuilder {
   private final IndexSchema schema;
   private Document doc;
-  private HashMap<String,String> map = new HashMap<String,String>();
+  private HashMap<String,String> map;
 
   public DocumentBuilder(IndexSchema schema) {
     this.schema = schema;
@@ -51,7 +51,7 @@
 
   public void startDoc() {
     doc = new Document();
-    map.clear();
+    map = new HashMap<String,String>();
   }
 
   protected void addSingleField(SchemaField sfield, String val, float boost) {
@@ -69,11 +69,6 @@
                   + ": first='" + oldValue + "' second='" + val + "'");
         }
       }
-      
-      if( doc == null ) {
-        throw new SolrException( SolrException.ErrorCode.SERVER_ERROR, 
-            "must call startDoc() before adding fields!" );
-      }
        
       // field.setBoost(boost);
       doc.add(field);
@@ -122,18 +117,21 @@
     
     // Check for all required fields -- Note, all fields with a
     // default value are defacto 'required' fields.  
-    List<String> missingFields = new ArrayList<String>( schema.getRequiredFields().size() );
+    List<String> missingFields = null;
     for (SchemaField field : schema.getRequiredFields()) {
       if (doc.getField(field.getName() ) == null) {
         if (field.getDefaultValue() != null) {
           doc.add( field.createField( field.getDefaultValue(), 1.0f ) );
         } else {
+          if (missingFields==null) {
+            missingFields = new ArrayList<String>(1);
+          }
           missingFields.add(field.getName());
         }
       }
     }
   
-    if (missingFields.size() > 0) {
+    if (missingFields != null) {
       StringBuilder builder = new StringBuilder();
       // add the uniqueKey if possible
       if( schema.getUniqueKeyField() != null ) {
@@ -181,7 +179,7 @@
       
       // Make sure it has the correct number
       Collection<Object> vals = doc.getFieldValues( name );
-      if( vals.size() > 1 && !sfield.multiValued() ) {
+      if(vals.size() > 1 && sfield!=null && !sfield.multiValued() ) {
         throw new SolrException( SolrException.ErrorCode.BAD_REQUEST,
             "ERROR: multiple values encountered for non multiValued field " + 
               sfield.getName() + ": " +vals.toString() );



Re: svn commit: r551060 - /lucene/solr/trunk/src/java/org/apache/solr/update/DocumentBuilder.java

Posted by Yonik Seeley <yo...@apache.org>.
On 6/27/07, Chris Hostetter <ho...@fucit.org> wrote:
>
> : check for null copyField source, delay some allocations
>
> i understanding delaying the allocation of missingFields, but why only a
> size of 1 once it is allocated?

Actually, it's an error case, so I don't really care about the speed
or memory consumption, as long as it only happens in that error case.
Could have just as easily left off the size.

-Yonik

Re: svn commit: r551060 - /lucene/solr/trunk/src/java/org/apache/solr/update/DocumentBuilder.java

Posted by Chris Hostetter <ho...@fucit.org>.
: check for null copyField source, delay some allocations

i understanding delaying the allocation of missingFields, but why only a
size of 1 once it is allocated?   (is this just an assumption thta
typically only a few fieldswill be missing?  would it make sense to just
switch to a LinkedLIst since missingFields is only ever used for
iteration?)


: -    List<String> missingFields = new ArrayList<String>( schema.getRequiredFields().size() );
: +    List<String> missingFields = null;
:      for (SchemaField field : schema.getRequiredFields()) {
:        if (doc.getField(field.getName() ) == null) {
:          if (field.getDefaultValue() != null) {
:            doc.add( field.createField( field.getDefaultValue(), 1.0f ) );
:          } else {
: +          if (missingFields==null) {
: +            missingFields = new ArrayList<String>(1);
: +          }


-Hoss