问题
I've collections of documents as like as below:
{
"_id" : ObjectId("55d4410544c96d6f6578f893"),
"executionProject" : "Project1",
"suiteList" : [
{
"suiteStatus" : "PASS",
}
],
"runEndTime" : ISODate("2015-08-19T08:40:47.049Z"),
"runStartTime" : ISODate("2015-08-19T08:40:37.621Z"),
"runStatus" : "PASS",
"__v" : 1
}
{
"_id" : ObjectId("55d44eb4c0422e7b8bffe76b"),
"executionProject" : "Project1",
"suiteList" : [
{
"suiteStatus" : "PASS",
}
],
"runEndTime" : ISODate("2015-08-19T09:39:13.528Z"),
"runStartTime" : ISODate("2015-08-19T09:39:00.406Z"),
"runStatus" : "PASS",
"__v" : 1
}
{
"_id" : ObjectId("55d44f0bc0422e7b8bffe76f"),
"executionProject" : "Project1",
"suiteList" : [
{
"suiteStatus" : "FAIL",
}
],
"runEndTime" : ISODate("2015-08-19T09:46:31.108Z"),
"runStartTime" : ISODate("2015-08-19T09:40:27.377Z"),
"runStatus" : "PASS",
"__v" : 1
}
{
"_id" : ObjectId("55d463d0c0422e7b8bffe789"),
"executionProject" : "Project2",
"suiteList" : [
{
"suiteStatus" : "FAIL"
},
{
"suiteStatus" : "PASS"
}
],
"runEndTime" : ISODate("2015-08-19T11:09:52.537Z"),
"runStartTime" : ISODate("2015-08-19T11:09:04.539Z"),
"runStatus" : "FAIL",
"__v" : 1
}
{
"_id" : ObjectId("55d464ebc0422e7b8bffe7c2"),
"executionProject" : "Project3",
"suiteList" : [
{
"suiteStatus" : "FAIL"
}
],
"runEndTime" : ISODate("2015-08-19T11:18:41.460Z"),
"runStartTime" : ISODate("2015-08-19T11:13:47.268Z"),
"runStatus" : "FAIL",
"__v" : 10
}
And I'm expecting output as follows:
[
{
"executionProject": "Project1",
"suite-pass": 0,
"suite-fail": 1,
"runEndTime": ISODate("2015-08-19T09:46:31.108Z")
},
{
"executionProject": "Project2",
"suite-pass": 1,
"suite-fail": 1,
"runEndTime": ISODate("2015-08-19T11:09:52.537Z")
},
{
"executionProject": "Project3",
"suite-pass": 0,
"suite-fail": 1,
"runEndTime": ISODate("2015-08-19T11:18:41.460Z")
},
]
I want to group by project and order by runEndTime and show the pass and fail counts of suiteList.
I know how to get the suite pass and fail counts of all runs, but not sure how can I group and order. Kindly help.
回答1:
The aggregation framework with the $cond operator seems to be what you are after:
Model.aggregate([
{ "$unwind": "$suiteList" },
{ "$group": {
"_id": "$executionProject",
"suite-pass": {
"$sum": {
"$cond": [
{ "$eq": [ "$suiteList.suiteStatus", "PASS" ] },
1,
0
]
}
},
"suite-fail": {
"$sum": {
"$cond": [
{ "$eq": [ "$suiteList.suiteStatus", "FAIL" ] },
1,
0
]
}
},
"runEndTime": { "$max": "$runEndTime" }
}},
{ "$sort": { "runEndTime": 1 }}
],function(err,result) {
});
This conditionally tests the values for either "PASS" or "FAIL" and returns these to the $sum accumulator under $group. In the same way you take the $max value for the relevant date per grouped documents.
The final thing is just to $sort by that date
If the "suiteList" in fact will only ever ccontain one "PASS" and one "FAIL" at maximum, then you can likely get away without even using $unwind on the array:
Model.aggregate(
[
{ "$group": {
"_id": "$executionProject",
"suite-pass": {
"$sum": {
"$cond": [
{ "$anyElementTrue": {
"$map": {
"input": "$suiteList",
"as": "suite",
"in": {
"$eq": [ "$$suite.suiteStatus", "PASS" ]
}
}
}},
1,
0
]
}
},
"suite-fail": {
"$sum": {
"$cond": [
{ "$anyElementTrue": {
"$map": {
"input": "$suiteList",
"as": "suite",
"in": {
"$eq": [ "$$suite.suiteStatus", "FAIL" ]
}
}
}},
1,
0
]
}
},
"runEndTime": { "$max": "$runEndTime" }
}},
{ "$sort": { "runEndTime": 1 }}
],
function(err,results) {
}
)
The $map and $anyElementTrue in there can similarly test the the conditions in the array matched where they should be counted. So as long as the "number of matches" per document does not matter, then this is fine.
On your data here that does meet those conditions, then both produce the same result:
{
"_id" : "Project1",
"suite-pass" : 2,
"suite-fail" : 1,
"runEndTime" : ISODate("2015-08-19T09:46:31.108Z")
}
{
"_id" : "Project2",
"suite-pass" : 1,
"suite-fail" : 1,
"runEndTime" : ISODate("2015-08-19T11:09:52.537Z")
}
{
"_id" : "Project3",
"suite-pass" : 0,
"suite-fail" : 1,
"runEndTime" : ISODate("2015-08-19T11:18:41.460Z")
}
To get the "last" items then simply $sort first and replace the $sum accumlators with the $last accumulator:
Model.aggregate(
[
{ "$sort": { "runEndTime": 1 } },
{ "$group": {
"_id": "$executionProject",
"suite-pass": {
"$last": {
"$cond": [
{ "$anyElementTrue": {
"$map": {
"input": "$suiteList",
"as": "suite",
"in": {
"$eq": [ "$$suite.suiteStatus", "PASS" ]
}
}
}},
1,
0
]
}
},
"suite-fail": {
"$last": {
"$cond": [
{ "$anyElementTrue": {
"$map": {
"input": "$suiteList",
"as": "suite",
"in": {
"$eq": [ "$$suite.suiteStatus", "FAIL" ]
}
}
}},
1,
0
]
}
},
"runEndTime": { "$last": "$runEndTime" }
}},
{ "$sort": { "runEndTime": 1 } }
],
function(err,results) {
}
);
Which produces:
{
"_id" : "Project1",
"suite-pass" : 0,
"suite-fail" : 1,
"runEndTime" : ISODate("2015-08-19T09:46:31.108Z")
}
{
"_id" : "Project2",
"suite-pass" : 1,
"suite-fail" : 1,
"runEndTime" : ISODate("2015-08-19T11:09:52.537Z")
}
{
"_id" : "Project3",
"suite-pass" : 0,
"suite-fail" : 1,
"runEndTime" : ISODate("2015-08-19T11:18:41.460Z")
}
来源：https://stackoverflow.com/questions/32175075/mongodb-group-by-element-and-show-the-sub-document-count-based-on-condition-and