부울 필드를 "참이 아님"(예 : 거짓 또는 존재하지 않음)으로 쿼리

부울 필드를 "참이 아님"(예 : 거짓 또는 존재하지 않음)으로 쿼리

---

나는 MongoDB 쿼리에서 매우 기본적인 것을 놓치고 있다고 확신합니다.이 간단한 조건을 얻을 수 없습니다.

이 컬렉션 고려

> db.tests.find()
{ "_id" : ObjectId("..."), "name" : "Test1" , "deleted" : true}
{ "_id" : ObjectId("..."), "name" : "Test2" , "deleted" : false}
{ "_id" : ObjectId("..."), "name" : "Test3" }

"삭제되지 않은"모든 항목을 쿼리하고 싶습니다.

"삭제됨"플래그가 true로 설정된 항목을 찾는 방법을 알고 있습니다.

> db.tests.find({deleted:true})
{ "_id" : ObjectId("..."), "name" : "Test1" , "deleted" : true}

그러나 그렇지 않은 모든 항목을 어떻게 찾 "deleted"습니까 (예 : 위의 쿼리를 부정하거나 다른 말로 "deleted"필드가 없거나 값이있는 항목)false

추측 해 본 것 (웃지 말아주세요 ...)

> db.tests.find({$not : {deleted: true}})

(결과를 반환하지 않음)

> db.tests.find({$not : {$eq:{deleted:true}}})

오류 : { "$ err": "잘못된 연산자 : $ eq", "code": 10068}

> db.tests.find({deleted:{$not: true}})

오류 : { "$ err": "$ not의 잘못된 사용", "code": 13041}

> db.tests.find({deleted:{$not: {$eq:true}}})

오류 : { "$ err": "$ not의 잘못된 사용", "code": 13034}

내가 무엇을 놓치고 있습니까?

---

db.tests.find({deleted: {$ne: true}})

어디 $ne를 의미합니다 "동일하지". ( mongodb 연산자에 대한 문서 )

---

완전성을 위해이 작업을 수행하는 또 다른 방법은 다음과 $in같습니다.

db.test.find({deleted: {$in: [null, false]}})

null배열에 포함 하면 deleted필드가 누락 된 문서를 가져옵니다 . 이 쿼리는 {deleted: 1}현재 2.6.6 MongoDB 릴리스 의 인덱스를 사용할 수 있습니다 .

---

JohnnyHK가 최고의 답을 가지고 있습니다. $in선택기 IMO 짧고 깨끗하다.

이것은 정확히 "거짓"또는 "존재하지 않음"을 테스트합니다. 그리고 색인화 될 수 있습니다.

db.tests.find({$or:[{deleted:false},{deleted:{$exists:false}}]})

인덱스를 사용한 예.

((function(){
    print("creating collection 'testx' and inserting 50 trues, 50 falses, 50 non-existents");
    db.testx.drop();
    db.testx.ensureIndex({deleted:1});
    for (var i=0;i<50;i++){
        db.testx.insert({i:i,deleted:false});
    };
    for (var i=0;i<50;i++){
        db.testx.insert({i:i,deleted:true});
    };
    for (var i=0;i<50;i++){
        db.testx.insert({i:i});
    };
    var res0 = db.testx.find().explain();
    var res1 = db.testx.find({deleted:false}).explain();
    var res2 = db.testx.find({deleted:true}).explain();
    var res3 = db.testx.find({deleted:{$exists:false}}).explain();
    var res4 = db.testx.find({$or:[{deleted:false},{deleted:{$exists:false}}]}).explain();
    var res5 = db.testx.find({$or:[{deleted:true},{deleted:{$exists:false}}]}).explain();
    var res6 = db.testx.find({deleted:{$in:[false,null]}}).explain();
    print("res0: all objects                      ("+res0["n"]+" found, "+res0["nscannedObjects"]+" scanned)");
    print("res1: deleted is false                 ("+res1["n"]+" found, "+res1["nscannedObjects"]+" scanned)");
    print("res2: deleted is true                  ("+res2["n"]+" found, "+res2["nscannedObjects"]+" scanned)");
    print("res3: deleted is non-existent          ("+res3["n"]+" found, "+res3["nscannedObjects"]+" scanned)");
    print("res4: deleted is false or non-existent ("+res4["n"]+" found, "+res4["nscannedObjects"]+" scanned)");
    print("res5: deleted is true or non-existent  ("+res5["n"]+" found, "+res5["nscannedObjects"]+" scanned)");
    print("res6: deleted is in [false,null]       ("+res5["n"]+" found, "+res5["nscannedObjects"]+" scanned)");
})())

이것은 인쇄되어야합니다

creating collection 'testx' and inserting 50 trues, 50 falses, 50 non-existents
res0: all objects                      (150 found, 150 scanned)
res1: deleted is false                 (50 found, 50 scanned)
res2: deleted is true                  (50 found, 50 scanned)
res3: deleted is non-existent          (50 found, 50 scanned)
res4: deleted is false or non-existent (100 found, 100 scanned)
res5: deleted is true or non-existent  (100 found, 100 scanned)
res6: deleted is in [false,null]       (100 found, 100 scanned)

---

For the case that someone needs this in an aggregation pipeline instead of find, this is what worked for me

db.getCollection('tests').aggregate([ 
  // ...previous operations...
  { $addFields: { "deleted_conclusion": { $cond: {
        if:{ $ne: [ "$deleted", false ]}, then: { $cond: [ "$deleted", ":TRUE", ":FALSY"]}, else: ":FALSE"
  }}}}
])

After adding the extra field you can go on with pipeline stages and have the information you miss

---

In case you are looking for mongoid syntax (I am using this in a rails app), this is what I came up with for a company's users:

2.3.1 :042 > accepted_consent = org.users.active.where(:accepted_terms_and_conditions => true).count
 => 553 
2.3.1 :043 > not_accepted_yet = org.users.active.where(:accepted_terms_and_conditions.ne => true).count
 => 6331 
2.3.1 :044 > 6331+553
 => 6884 
2.3.1 :045 > org.users.active.count
 => 6884 

참고URL : https://stackoverflow.com/questions/18837486/query-for-boolean-field-as-not-true-e-g-either-false-or-non-existent