@@ -42,16 +42,100 @@ fields:
4242
4343 :method:`db.collection.createIndex()`
4444
45- Index Hints for Time Series Collections
46- ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
45+ .. _timeseries-secondary-index-sort-performance:
4746
48- To force MongoDB to use a specific index when querying a :term:`time
49- series collection`, use :method:`cursor.hint()` with an index name:
47+ Use Secondary Indexes to Improve Sort Performance
48+ -------------------------------------------------
49+
50+ Sort operations on the ``timeField`` and ``metaField`` can use secondary
51+ indexes on those fields to improve performance.
52+
53+ For example, the following ``sensorData`` collection contains
54+ temperature readings:
55+
56+ .. code-block:: javascript
57+
58+ db.sensorData.insertMany( [
59+ {
60+ "metadata": { "sensorId": 5578, "type": "temperature" },
61+ "timestamp": ISODate("2022-01-15T00:00:00.000Z"),
62+ "temperatureReading": 12
63+ },
64+ {
65+ "metadata": { "sensorId": 5578, "type": "temperature" },
66+ "timestamp": ISODate("2022-01-15T04:00:00.000Z"),
67+ "temperatureReading": 11
68+ },
69+ {
70+ "metadata": { "sensorId": 5579, "type": "temperature" },
71+ "timestamp": ISODate("2022-01-15T08:00:00.000Z"),
72+ "temperatureReading": 9
73+ }
74+ ] )
75+
76+ The following command creates a compound ascending secondary index on
77+ the ``timestamp`` and ``metadata.sensorId`` fields:
78+
79+ .. code-block:: javascript
80+
81+ db.sensorData.createIndex(
82+ { "timestamp": 1, "metadata.sensorId": 1 }
83+ )
84+
85+ The following sort operation on the ``timestamp`` field uses the index
86+ to improve performance:
87+
88+ .. code-block:: javascript
89+
90+ db.sensorData.find().sort( { "timestamp": 1 } )
91+
92+ To confirm that the sort operation used the index, run the operation
93+ again with the ``.explain()`` option:
94+
95+ .. code-block:: javascript
96+
97+ db.sensorData.find().sort( { "timestamp": 1 } ).explain()
98+
99+ The ``winningPlan.queryPlan.inputStage.stage`` is ``IXSCAN``, which
100+ indicates that the index was used. For more information on explain plan
101+ output, see :ref:`explain-results`.
102+
103+ Specify Index Hints for Time Series Collections
104+ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
105+
106+ Index hints cause MongoDB to use a specific index for a query. Some
107+ operations on time series collections can only take advantage of an
108+ index if that index is specified in a hint.
109+
110+ For example, the following query causes MongoDB to use the
111+ ``timestamp_1_metadata.sensorId_1`` index:
50112
51113.. code-block:: javascript
52114
53- db.weather24h.find({ "metadata.sensorId": 1235 }).hint("metadata.sensorId_1_timestamp_1")
115+ db.sensorData.find( { "metadata.sensorId": 5578 } ).hint( "timestamp_1_metadata.sensorId_1" )
116+
117+ On a time series collection, you can specify hints using either the
118+ index name or the index key pattern. To get the names of the indexes on
119+ a collection, use the :method:`db.collection.getIndexes()` method.
120+
121+ .. _timeseries-add-secondary-index-mongodb-6.0:
122+
123+ Time Series Secondary Indexes
124+ -----------------------------
125+
126+ Starting in MongoDB 6.0 (and 5.0.16):
127+
128+ - You can add a :doc:`compound index </core/index-compound>` on time,
129+ metadata, or measurement fields.
130+
131+ - You can use the :query:`$or`, :query:`$in`, and :query:`$geoWithin`
132+ operators with :doc:`partial indexes </core/index-partial>` on a time
133+ series collection.
134+
135+ - You can add :doc:`partial </core/index-partial>` and :doc:`2dsphere
136+ </core/2dsphere>` indexes on any field in a :ref:`time series
137+ collection <manual-timeseries-collection>`.
138+
139+ .. note::
54140
55- On a time series collection, you can only specify hints using the index
56- name, not the index key pattern. To get the names of the indexes on a
57- collection, use the :method:`db.collection.getIndexes()` method.
141+ .. include:: /includes/time-series-secondary-indexes-downgrade-FCV.rst
0 commit comments