From cfe1059092fdface253c359739079a3434b1b2a0 Mon Sep 17 00:00:00 2001
From: Yuya Ebihara <yuya.ebihara@starburstdata.com>
Date: Thu, 24 Feb 2022 12:41:12 +0900
Subject: [PATCH 1/2] Document guess field type mapping in MongoDB

Additionally, improve the sentence how to handle wrong table
definitions.
---
 docs/src/main/sphinx/connector/mongodb.rst | 22 +++++++++++++++++++++-
 1 file changed, 21 insertions(+), 1 deletion(-)

diff --git a/docs/src/main/sphinx/connector/mongodb.rst b/docs/src/main/sphinx/connector/mongodb.rst
index 04bf342606ce..5c86340373b9 100644
--- a/docs/src/main/sphinx/connector/mongodb.rst
+++ b/docs/src/main/sphinx/connector/mongodb.rst
@@ -84,7 +84,27 @@ This property is required; there is no default. A connection url or seeds must b
 
 As MongoDB is a document database, there is no fixed schema information in the system. So a special collection in each MongoDB database should define the schema of all tables. Please refer the :ref:`table-definition-label` section for the details.
 
-At startup, this connector tries guessing fields' types, but it might not be correct for your collection. In that case, you need to modify it manually. ``CREATE TABLE`` and ``CREATE TABLE AS SELECT`` will create an entry for you.
+At startup, the connector tries to guess the data type of fields based on the mapping in the following table.
+
+================== ================ ================================================
+MongoDB            Trino            Notes
+================== ================ ================================================
+``Boolean``        ``BOOLEAN``
+``Int32``          ``BIGINT``
+``Int64``          ``BIGINT``
+``Double``         ``DOUBLE``
+``Date``           ``TIMESTAMP(3)``
+``String``         ``VARCHAR``
+``Binary``         ``VARBINARY``
+``ObjectId``       ``ObjectId``
+``Object``         ``ROW``
+``Array``          ``ARRAY``        Map to ``ROW`` if the element type is not unique
+``DBRef``          ``ROW``
+================== ================ ================================================
+
+The initial guess can be incorrect for your specific collection. In that case, you need to modify it manually. Please refer the :ref:`table-definition-label` section for the details.
+
+Creating new tables using ``CREATE TABLE`` and ``CREATE TABLE AS SELECT`` automatically create an entry for you.
 
 This property is optional; the default is ``_schema``.
 

From 131976ca0fe61863638498d354ed60105b155cf2 Mon Sep 17 00:00:00 2001
From: Yuya Ebihara <yuya.ebihara@starburstdata.com>
Date: Thu, 24 Feb 2022 12:40:40 +0900
Subject: [PATCH 2/2] Add test to guess field types in MongoDB

---
 .../mongodb/BaseMongoConnectorTest.java       | 33 +++++++++++++++++++
 1 file changed, 33 insertions(+)

diff --git a/plugin/trino-mongodb/src/test/java/io/trino/plugin/mongodb/BaseMongoConnectorTest.java b/plugin/trino-mongodb/src/test/java/io/trino/plugin/mongodb/BaseMongoConnectorTest.java
index b41b0ad93344..bd9c87608931 100644
--- a/plugin/trino-mongodb/src/test/java/io/trino/plugin/mongodb/BaseMongoConnectorTest.java
+++ b/plugin/trino-mongodb/src/test/java/io/trino/plugin/mongodb/BaseMongoConnectorTest.java
@@ -104,6 +104,39 @@ public void testSortItemsReflectedInExplain()
                 "TopNPartial\\[5 by \\(nationkey DESC");
     }
 
+    @Test(dataProvider = "guessFieldTypesProvider")
+    public void testGuessFieldTypes(String mongoValue, String trinoValue)
+    {
+        Document document = Document.parse(format("{\"test\":%s}", mongoValue));
+
+        assertUpdate("DROP TABLE IF EXISTS test.test_guess_field_type");
+        client.getDatabase("test").getCollection("test_guess_field_type").insertOne(document);
+
+        assertThat(query("SELECT test FROM test.test_guess_field_type"))
+                .matches("SELECT " + trinoValue);
+
+        assertUpdate("DROP TABLE test.test_guess_field_type");
+    }
+
+    @DataProvider
+    public Object[][] guessFieldTypesProvider()
+    {
+        return new Object[][] {
+                {"true", "true"}, // boolean -> boolean
+                {"2147483647", "bigint '2147483647'"}, // int32 -> bigint
+                {"{\"$numberLong\": \"9223372036854775807\"}", "9223372036854775807"}, // int64 -> bigint
+                {"1.23", "double '1.23'"}, // double -> double
+                {"{\"$date\": \"1970-01-01T00:00:00.000Z\"}", "timestamp '1970-01-01 00:00:00.000'"}, // date -> timestamp(3)
+                {"'String type'", "varchar 'String type'"}, // string -> varchar
+                {"{$binary: \"\",\"$type\": \"0\"}", "to_utf8('')"}, // binary -> varbinary
+                {"{\"$oid\": \"6216f0c6c432d45190f25e7c\"}", "ObjectId('6216f0c6c432d45190f25e7c')"}, // objectid -> objectid
+                {"[1]", "array[bigint '1']"}, // array with single type -> array
+                {"{\"field\": \"object\"}", "CAST(row('object') AS row(field varchar))"}, // object -> row
+                {"[9, \"test\"]", "CAST(row(9, 'test') AS row(_pos1 bigint, _pos2 varchar))"}, // array with multiple types -> row
+                {"{\"$ref\":\"test_ref\",\"$id\":ObjectId(\"4e3f33de6266b5845052c02c\"),\"$db\":\"test_db\"}", "CAST(row('test_db', 'test_ref', ObjectId('4e3f33de6266b5845052c02c')) AS row(databasename varchar, collectionname varchar, id ObjectId))"}, // dbref -> row
+        };
+    }
+
     @Test
     public void createTableWithEveryType()
     {