Added tests for surrogate-based indexed NL fuzzy joins, where the top-level join is unecessary and optimized away.
git-svn-id: https://asterixdb.googlecode.com/svn/branches/asterix_fuzzy_perf@891 eaa15691-b419-025a-1212-ee371bd00084
diff --git a/asterix-app/src/test/resources/optimizerts/queries/inverted-index-join-noeqjoin/ngram-fuzzyeq-edit-distance.aql b/asterix-app/src/test/resources/optimizerts/queries/inverted-index-join-noeqjoin/ngram-fuzzyeq-edit-distance.aql
new file mode 100644
index 0000000..87ebe4a
--- /dev/null
+++ b/asterix-app/src/test/resources/optimizerts/queries/inverted-index-join-noeqjoin/ngram-fuzzyeq-edit-distance.aql
@@ -0,0 +1,42 @@
+/*
+ * Description : Fuzzy joins two datasets, DBLP and CSX, based on ~= using edit distance of their authors.
+ * DBLP has a 3-gram index on authors, and we expect the join to be transformed into an indexed nested-loop join.
+ * We expect the top-level equi join introduced because of surrogate optimization to be removed, since it is not necessary.
+ * Success : Yes
+ */
+
+drop dataverse test if exists;
+create dataverse test;
+use dataverse test;
+
+create type DBLPType as closed {
+ id: int32,
+ dblpid: string,
+ title: string,
+ authors: string,
+ misc: string
+}
+
+create type CSXType as closed {
+ id: int32,
+ csxid: string,
+ title: string,
+ authors: string,
+ misc: string
+}
+
+create dataset DBLP(DBLPType) partitioned by key id;
+
+create dataset CSX(CSXType) partitioned by key id;
+
+create index ngram_index on CSX(authors) type ngram(3);
+
+write output to nc1:"rttest/inverted-index-join-noeqjoin_ngram-fuzzyeq-edit-distance.adm";
+
+set simfunction 'edit-distance';
+set simthreshold '3';
+
+for $a in dataset('DBLP')
+for $b in dataset('CSX')
+where $a.authors ~= $b.authors and $a.id < $b.id
+return {"aauthors": $a.authors, "bauthors": $b.authors}
\ No newline at end of file