Hello guys.I am using bucket pattern for time-series.My data look like this:
{'_id': ObjectId('6068da8878fa2e568c42c7f1'),
'first': datetime.datetime(2018, 1, 24, 14, 5),
'last': datetime.datetime(2018, 1, 24, 15, 5),
'maxid13': 12.5,
'minid13': 7.5,
'nsamples': 13,
'samples': [{'c14': 'C',
'id1': 3758.0,
'id10': 0.0,
'id11': 274.0,
'id12': 0.0,
'id13': 7.5,
'id15': 0.0,
'id16': 73.0,
'id17': 0.0,
'id18': 0.342,
'id19': 6.3,
'id20': 1206.0,
'id21': 0.0,
'id22': 0.87,
'id23': 0.0,
'id6': 2.0,
'id7': -79.09,
'id8': 35.97,
'id9': 5.8,
'timestamp1': datetime.datetime(2018, 1, 24, 14, 5),
'timestamp2': datetime.datetime(2018, 1, 24, 9, 5)},
{'c14': 'C',
'id1': 3758.0,
'id10': 0.0,
'id11': 288.0,
'id12': 0.0,
'id13': 8.4,
'id15': 0.0,
'id16': 71.0,
'id17': 0.0,
'id18': 0.342,
'id19': 6.3,
'id20': 1207.0,
'id21': 0.0,
'id22': 0.69,
'id23': 0.0,
'id6': 2.0,
'id7': -79.09,
'id8': 35.97,
'id9': 6.2,
'timestamp1': datetime.datetime(2018, 1, 24, 14, 10),
'timestamp2': datetime.datetime(2018, 1, 24, 9, 10)},
.
.
.
.
In the most cases in the match
stage i use first
with samples.id13
andsamples.id9
and in some other queries i use samples.timestamp1
with samples.id13
and samples.id9
,and in sort
stage sometimes i use first
.Should i create 4 single index on them or a compound(with the correct order)?