Github user moshebla commented on a diff in the pull request:
https://github.com/apache/lucene-solr/pull/395#discussion_r195118056
--- Diff: solr/core/src/java/org/apache/solr/handler/loader/JsonLoader.java
---
@@ -703,6 +703,10 @@ private SolrInputDocument generateExtendedValueDoc(int
ev) throws IOException {
String label = parser.getString();
SolrInputField sif = new SolrInputField(label);
parseFieldValue(sif);
+ // pulling out the pieces may seem weird, but it's because
--- End diff --
Skimming through the code on master, it looks as if the map parsing was
done here before only partially. Currently it seems like we either duplicate
the map parsing part, or the document parsing part. If we go with the map part,
we would build an intermediate map and then parse it to a document instead of
streaming it using the parser. I'm not quite sure if using buildDoc would be as
beneficial, since in either case, we have to duplicate some part.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org
For additional commands, e-mail: dev-h...@lucene.apache.org
