I have this code in MySQL
corresponding to a simple car database
SELECT
"Cost" AS "Metric",
AVG(car_cost) As "Avg",
MIN(car_cost) As "Min",
MAX(car_cost) As "Max",
"Dollars" As "Unit"
from db.car_data
union SELECT
'Mileage',
AVG(car_mileage),
MIN(car_mileage),
MAX(car_mileage),
'Miles'
from db.car_data
That returns this table
Metric | Avg | Min | Max |
---|---|---|---|
Cost | 2345 | 5674 | 53456 |
Mileage | 49023 | 53 | 178003 |
But, when recreate this in a MongoDB
equivalent I run into an issue
db.car_data.aggregate([
{"$group":
{
_id: null,
Avg: {
$avg: { "$car_cost" }
},
Min: {
$min: { "$car_cost" }
},
Max: {
$max: { "$car_cost" }
}
}
},
{"$group":
{
_id: null,
Avg: {
$avg: { car_mileage }
},
Min: {
$min: { car_mileage }
},
Max: {
$max: { car_mileage }
}
}
}
]);
Actual Output
[
{
"_id": null,
"Avg": null,
"Min": null,
"Max": null
}
]
Expected Output
[
{
"_id": null,
"Avg": 2345,
"Min": 5674,
"Max": 53456
},
{
"_id": null,
"Avg": 49023,
"Min": 53,
"Max": 178003
}
]
Am I doing something wrong or is there a better way to do this with aggregation in Mongo? They each work in Mongo on their own...
CodePudding user response:
Query
- you have 1 collection in 1 database, and you want 6 accumulators
- you can seperate them in groups of 3 or even in 2 documents to get the exact SQL like result
*we have $facet
$unionWith
etc but you dont need those, no need to do multiple aggregations, with only 1 you can get all you need.
aggregate(
[ {
"$group" : {
"_id" : null,
"avg_cost" : {
"$avg" : "$car_cost"
},
"min_cost" : {
"$min" : "$car_cost"
},
"max_cost" : {
"$max" : "$car_cost"
},
"avg_mileage" : {
"$avg" : "$car_mileage"
},
"min_mileage" : {
"$min" : "$car_mileage"
},
"max_mileage" : {
"$max" : "$car_mileage"
}
}
}, {
"$set" : {
"array" : [ {
"metric" : "cost",
"avg" : "$avg_cost",
"min" : "$min_cost",
"max" : "$max_cost"
}, {
"metric" : "mileage",
"avg" : "$avg_mileage",
"min" : "$min_mileage",
"max" : "$max_mileage"
} ]
}
}, {
"$unwind" : {
"path" : "$array"
}
}, {
"$replaceRoot" : {
"newRoot" : "$array"
}
} ]
)