mongodb count two condition one time? - mongodb

Question

I need to count two query like below:
1. data[k]['success'] = common.find({'url': {"$regex": k}, 'result.titile':{'$ne':''}}).count()
2. data[k]['fail'] = common.find({'url': {"$regex": k}, 'result.titile':''}).count()
I think it would be more efficient if mongodb can work like below:
result = common.find({'url': {"$regex": k})
count1 = result.find({'result.titile':{'$ne':''}})
count2 = result.count() - count1
//result do not have find or count method, just for example
Two count are basing same search condition{'url': {"$regex": k}, splited by {'result.titile':{'$ne':''}} or not.
Is there some build-in way to do these without writing custom js?

Answer 1

The async method would be the preferred one if at all your client supports it.
You could also aggregate as below:
$match the docs which have the urls.
$group by the _id as null, and take the $sum of all documents. We need those documents, to get the sum of those which do not have a title, so just accumulate them using the $push operator.
$unwind the documents.
$match those which do not have a title.
$group, and get the $sum.
$project the desired result.
sample code:
db.t.aggregate([
{$match:{"url":{"$regex":k}}},
{$group:{"_id":null,
"count_of_url_matching_docs":{$sum:1},
"docs":{$push:"$$ROOT"}}},
{$unwind:"$docs"},
{$match:{"docs.result.titile":{$ne:""}}},
{$group:{"_id":null,
"count_of_url_matching_docs":{$first:"$count_of_url_matching_docs"},
"count_of_docs_with_titles":{$sum:1}}},
{$project:{"_id":0,
"count_of_docs_with_titles":"$count_of_docs_with_titles",
"count_difference":{$subtract:[
"$count_of_url_matching_docs",
"$count_of_docs_with_titles"]}}}
])
Test data:
db.t.insert([
{"url":"s","result":{"titile":1}},
{"url":"s","result":{"titile":""}},
{"url":"s","result":{"titile":""}},
{"url":"s","result":{"titile":2}}
])
Test Result:
{ "count_of_docs_with_titles" : 2, "count_difference" : 2 }

Answer 2

Use .aggregate() with a conditional key for grouping via $cond:
common.aggregate([
{ "$match": { "url": { "$regex": k } } },
{ "$group": {
"_id": {
"$cond": {
"if": { "$ne": [ "$result.title", "" ] },
"then": "success",
"else": "fail"
}
},
"count": { "$sum": 1 }
}}
])
However it is actually more efficient to run both queries in parallel if your environment supports it, such as with nodejs
async.parallel(
[
function(callback) {
common.count({
"url": { "$regex": k },
"result.title": { "$ne": "" }
}, function(err,count) {
callback(err,{ "success": count });
});
},
function(callback) {
common.count({
"url": { "$regex": k },
"result.title": ""
}, function(err,count) {
callback(err,{ "fail": count });
});
}
],
function(err,results) {
if (err) throw err;
console.log(results);
}
)
Which makes sense really since each item is not being tested and each result can actually run on the server at the same time.