Fix encoding for OpenSearch count pushdown

trinodb · Sep 19, 2024 · 3c1b11c · 3c1b11c
1 parent 52821d2
commit 3c1b11c
Show file tree

Hide file tree

Showing 2 changed files with 33 additions and 1 deletion.
diff --git a/...in/trino-opensearch/src/main/java/io/trino/plugin/opensearch/client/OpenSearchClient.java b/...in/trino-opensearch/src/main/java/io/trino/plugin/opensearch/client/OpenSearchClient.java
@@ -667,7 +667,7 @@ public long count(String index, int shard, QueryBuilder query)
                                 "GET",
                                 format("/%s/_count?preference=_shards:%s", index, shard),
                                 ImmutableMap.of(),
-                                new StringEntity(sourceBuilder.toString()),
+                                new StringEntity(sourceBuilder.toString(), UTF_8),
                                 new BasicHeader("Content-Type", "application/json"));
             }
             catch (ResponseException e) {

diff --git a/...rino-opensearch/src/test/java/io/trino/plugin/opensearch/BaseOpenSearchConnectorTest.java b/...rino-opensearch/src/test/java/io/trino/plugin/opensearch/BaseOpenSearchConnectorTest.java
@@ -1583,6 +1583,38 @@ public void testFilters()
         assertQuery("SELECT count(*) FROM filter_pushdown WHERE ipv6_column = IPADDRESS '2001:db8::1:0:0:1'", "VALUES 1");
     }
 
+    @Test
+    public void testFiltersCharset()
+            throws IOException
+    {
+        String indexName = "filter_charset_pushdown";
+
+        @Language("JSON")
+        String mappings = """
+                {
+                  "properties": {
+                    "keyword_column":   { "type": "keyword" },
+                    "text_column":      { "type": "text" }
+                  }
+                }
+                """;
+
+        createIndex(indexName, mappings);
+
+        index(indexName, ImmutableMap.<String, Object>builder()
+                .put("keyword_column", "Türkiye")
+                .put("text_column", "Türkiye")
+                .buildOrThrow());
+
+        assertQuery("SELECT count(*) FROM filter_charset_pushdown WHERE keyword_column = 'Türkiye'", "VALUES 1");
+        assertQuery("SELECT count(*) FROM filter_charset_pushdown WHERE keyword_column = 'bar'", "VALUES 0");
+        assertQuery("SELECT count(*) FROM filter_charset_pushdown WHERE text_column = 'Türkiye'", "VALUES 1");
+        assertQuery("SELECT count(*) FROM filter_charset_pushdown WHERE text_column = 'some'", "VALUES 0");
+
+        assertQuery("SELECT keyword_column FROM filter_charset_pushdown WHERE keyword_column = 'Türkiye'", "VALUES ('Türkiye')");
+        assertQuery("SELECT text_column FROM filter_charset_pushdown WHERE text_column = 'Türkiye'", "VALUES ('Türkiye')");
+    }
+
     @Test
     public void testLimitPushdown()
             throws IOException