I need to ensure combination of some fields to be unique , so i found out we can use indexes to achieve it , like :
db.user.ensureIndex({'fname':1,'lname':1},{unique:true})
However in my scenario i allow the user to add email or phone or even both . So how do i achieve the uniqueness in this case . I tried like setting 2 indexed like
db.user.ensureIndex({'fname':1,'lname':1,'email':1},{unique:true})
And
db.user.ensureIndex({'fname':1,'lname':1,'mobile':1},{unique:true})
this dint work when i try to add a document with email unique , like below
> db.user.insert({fname:'tony',lname:'stark',email:'aa#gmail.com'})
WriteResult({ "nInserted" : 1 })
> db.user.insert({fname:'tony',lname:'stark',email:'xyz#gmail.com'})
WriteResult({
"nInserted" : 0,
"writeError" : {
"code" : 11000,
"errmsg" : "E11000 duplicate key error index: temp.user.$fname_1_lname_1_mobile_1 dup key: { : \"tony\", : \"stark\", : null }"
}
})
I guess the error is because of the second index ,since both the inserts have null value for the mobile field . I then tried this
db.user.ensureIndex({'fname':1,'lname':1,'email':1,'mobile':1},{unique:true})
And that dint help ,is there any way to do this with mongodb or check the uniqueness with code otherwise .
Update
I need to ensure uniqueness in this case too :
db.user.insert({fname:'tony',lname:'stark',email:'abc#gmail.com'})
<<insert>>
db.user.insert({fname:'tony',lname:'stark',email:'abc#gmail.com',mobile:554545435})
<<dont insert>>
Update 2
Even Unique partial Index didn't help
> db.users.createIndex({ "fname": 1, "lname": 1, "email":1 },
... { unique: true, partialFilterExpression:{
... email: { $exists: true, $gt : { $type : 10 } } } })
{
"createdCollectionAutomatically" : true,
"numIndexesBefore" : 1,
"numIndexesAfter" : 2,
"ok" : 1
}
> db.users.createIndex({ "fname": 1, "lname": 1, "mobile":1 },
... { unique: true, partialFilterExpression:{
... mobile: { $exists: true, $gt : { $type : 10 } } } })
{
"createdCollectionAutomatically" : false,
"numIndexesBefore" : 2,
"numIndexesAfter" : 3,
"ok" : 1
}
Here is the query :
>db.users.insert({fname:"tony",lname:"stark",mobile:"123",email:"123#gmail.com"})
WriteResult({ "nInserted" : 1 })
>db.users.insert({fname:"tony",lname:"stark",mobile:"123",email:"123#gmail.com"})
WriteResult({ "nInserted" : 1 })
Mongo version - v3.2.0
I have no experience with mongodb much neither have used indexes in other db's , i guess it's kind of noob question .Thanks in advance
Related
Say I have the fields a and b. I want to have a compound uniqueness where if a: 1, b: 2, I would not be able to do a: 2, b: 1.
The reason I want this is because I'm making a "friends list" kind of collection, where if a is connected to b, then it's automatically the reverse as well.
is this possible on a schema level or do I need to do queries to check.
If you don't need to differentiate between requester and requestee, you could sort the values before saving or querying so that your two fields a and b have a predictable order for any pair of friend IDs (and you can take advantage of the unique index constraint).
For example, using the mongo shell:
Create a helper function to return friend pairs in predictable order:
function friendpair (friend1, friend2) {
if ( friend1 < friend2) {
return ({a: friend1, b: friend2})
} else {
return ({a: friend2, b: friend1})
}
}
Add a compound unique index:
> db.friends.createIndex({a:1, b:1}, {unique: true});
{
"createdCollectionAutomatically" : true,
"numIndexesBefore" : 1,
"numIndexesAfter" : 2,
"ok" : 1
}
Insert unique pairs (should work)
> db.friends.insert(friendpair(1,2))
WriteResult({ "nInserted" : 1 })
> db.friends.insert(friendpair(1,3))
WriteResult({ "nInserted" : 1 })
Insert non-unique pair (should return duplicate key error):
> db.friends.insert(friendpair(2,1))
WriteResult({
"nInserted" : 0,
"writeError" : {
"code" : 11000,
"errmsg" : "E11000 duplicate key error collection: test.friends index: a_1_b_1 dup key: { : 1.0, : 2.0 }"
}
})
Search should work in either order:
db.friends.find(friendpair(3,1)).pretty()
{ "_id" : ObjectId("5bc80ed11466009f3b56fa52"), "a" : 1, "b" : 3 }
db.friends.find(friendpair(1,3)).pretty()
{ "_id" : ObjectId("5bc80ed11466009f3b56fa52"), "a" : 1, "b" : 3 }
Instead of handling duplicate key errors or insert versus update, you could also use findAndModify with an upsert since this is expected to be a unique pair:
> var pair = friendpair(2,1)
> db.friends.findAndModify({
query: pair,
update: {
$set: {
a : pair.a,
b : pair.b
},
$setOnInsert: { status: 'pending' },
},
upsert: true
})
{
"_id" : ObjectId("5bc81722ce51da0e4118c92f"),
"a" : 1,
"b" : 2,
"status" : "pending"
}
Doesn't seem like you can do a unique on the entire array's values so I'm doing a kind of work around. I'm using the $jsonSchema as follows:
{
$jsonSchema:
{
bsonType:
"object",
required:
[
"status",
"users"
],
properties:
{
status:
{
enum:
[
"pending",
"accepted"
],
bsonType:
"string"
},
users:
{
bsonType:
"array",
description:
"references two user_id",
items:
{
bsonType:
"objectId"
},
maxItems:
2,
minItems:
2,
},
}
}
}
then I will use $all to find the connected users, e.g.
db.collection.find( { users: { $all: [ ObjectId1, ObjectId2 ] } } )
I want to create an index so that my DB does not allow me to insert documents whose value for the key lema is already present in some document of the DB. I did this:
db.version()
3.0.14
> db.rae.ensureIndex({"lema":1, unique: true})
{
"createdCollectionAutomatically" : false,
"numIndexesBefore" : 2,
"numIndexesAfter" : 2,
"note" : "all indexes already exist",
"ok" : 1
}
> db.rae.insert({"lema":"a"})
WriteResult({ "nInserted" : 1 })
> db.rae.insert({"lema":"a"})
WriteResult({ "nInserted" : 1 })
> db.rae.insert({"lema":"a"})
WriteResult({ "nInserted" : 1 })
> db.rae.find()
{ "_id" : ObjectId("591a0ce372329f3162a314cc"), "lema" : "a" }
{ "_id" : ObjectId("591a0ce472329f3162a314cd"), "lema" : "a" }
{ "_id" : ObjectId("591a0ce572329f3162a314ce"), "lema" : "a" }
Clearly the DB is letting me insert documents whose values of lema are all a. How can I fix this? Thanks a lot
From Stennie's comment:
I should use createIndex instead of ensureIndex. Also I had a mistake, I should use db.rae.createIndex({"lema":1}, { unique: true}).
I need to store three id:s in a document but they can only occur once. For example the document below can only occur once in this collection:
{
"user": ObjectId("j8uwh902w5489"),
"comment": ObjectId("09890583457jkjsf4"),
"whatever": ObjectId("j8uwh902w5489")
}
How do I make sure this will be a unique document in MongoDB?
Well you can use MongoDB's Unique Compound Index
db.users.createIndex( { user: 1, comment: 1, whatever: 1 }, { unique: true } )
A few Cases:
> db.users.insert({user: "A", comment: "B", whatever: "C"})
WriteResult({ "nInserted" : 1 })
> db.users.insert({user: "A", comment: "C", whatever: "B"})
WriteResult({ "nInserted" : 1 })
> db.users.insert({user: "A", comment: "B", whatever: "C"})
> WriteResult({
"nInserted" : 0,
"writeError" : {
"code" : 11000,
"errmsg" : "insertDocument :: caused by :: 11000 E11000 duplicate key error index: test.users.$user_1_comment_1_whatev
er_1 dup key: { : \"A\", : \"B\", : \"C\" }"
}
})
I know I have used strings here. But something similar might also be possible with ObjectIds. Please do give it a try.
I have a collection where I will have a _children attribute like this:
{
_children: {
videoTags: [ { id: '1', name: 'one'}, { id: '2', name: 'two'} ],
},
a: 10
}
Since I WILL search in videoTags, I create an index as such:
> db.test4.createIndex({ "_children.videosTags.id" : 1 }, { "unique" : true } );
{
"createdCollectionAutomatically" : false,
"numIndexesBefore" : 1,
"numIndexesAfter" : 2,
"ok" : 1
}
Trouble is, I can no longer add anything to that table since I get a duplicate index error. Here is how to reproduce it:
Step 1: insert to a collection
db.test4.insert({a:20})
WriteResult({ "nInserted" : 1 })
Step 2: make the index
db.test4.createIndex({ "_children.videosTags.id" : 1 }, { "unique" : true } );
{
"createdCollectionAutomatically" : false,
"numIndexesBefore" : 1,
"numIndexesAfter" : 2,
"ok" : 1
}
Step 3: try to insert again
db.test4.insert({a:30})
WriteResult({
"nInserted" : 0,
"writeError" : {
"code" : 11000,
"errmsg" : "insertDocument :: caused by :: 11000 E11000 duplicate key error index: wonder_1.test4.$_children.videosTags.id_1 dup key: { : null }"
}
})
I think the issue here is that there is already a record where _children.videoTags.id is not defined.
However, what I expected was a behaviour where if videoTags.id was specified, it needed to be unique. Instead, an empty one is considered a "taken" key.
What am I doing that is stupidly wrong?
This will work if you don't set unique as true, but I have the feeling I need to fix it for real...
There could be two reasons.
There could be other documents exists in collection with same _children.videosTags.id
It's quite possible that more than one document may have missing _children.videosTags.id" or having null value.
As you are creating unique key, null or empty values are give you tough time. Solution is either create sparse index and if your MongoDB version is 3.2+, create partial index. See documentation for partial indexes.
I have an post whose field postStatus I want to set as either true false, or null. I was wondering whether mongo would be able to index null values.
Also, is there any difference if I choose to do null, true and false vs using -1,0,1 (or any 3 ints)?
thanks
Indexing of null
Yes, null values are indexed, as can be proven if you try to add a second null value to a unique index:
connecting to: test
> db.idxtest.createIndex({a:1},{unique:true})
{
"createdCollectionAutomatically" : true,
"numIndexesBefore" : 1,
"numIndexesAfter" : 2,
"ok" : 1
}
> db.idxtest.insert({a:"foo"})
WriteResult({ "nInserted" : 1 })
> db.idxtest.insert({b:"bar"})
WriteResult({ "nInserted" : 1 })
> db.idxtest.insert({c:"baz"})
WriteResult({
"nInserted" : 0,
"writeError" : {
"code" : 11000,
"errmsg" : "E11000 duplicate key error index: test.idxtest.$a_1 dup key: { : null }"
}
})
Another way to demonstrate that null values matter is by sorting on an indexed field for which documents contain null values:
> db.idxtest2.createIndex({a:1})
{
"createdCollectionAutomatically" : true,
"numIndexesBefore" : 1,
"numIndexesAfter" : 2,
"ok" : 1
}
> db.idxtest2.insert({a:1})
WriteResult({ "nInserted" : 1 })
> db.idxtest.insert({a:2})
WriteResult({ "nInserted" : 1 })
> db.idxtest2.insert({b:2})
WriteResult({ "nInserted" : 1 })
> db.idxtest2.insert({a:null})
WriteResult({ "nInserted" : 1 })
> db.idxtest2.find().sort({a:1})
{ "_id" : ObjectId("56786a93ada44c7ffcd38f9f"), "b" : 2 }
{ "_id" : ObjectId("56786aa1ada44c7ffcd38fa0"), "a" : null }
{ "_id" : ObjectId("56786a65ada44c7ffcd38f9d"), "a" : 1 }
> db.idxtest2.find().sort({a:-1})
{ "_id" : ObjectId("56786a65ada44c7ffcd38f9d"), "a" : 1 }
{ "_id" : ObjectId("56786aa1ada44c7ffcd38fa0"), "a" : null }
{ "_id" : ObjectId("56786a93ada44c7ffcd38f9f"), "b" : 2 }
Note that implicit null values have precedence; so to say, they are "more" null since they do not even have the according key.
On Booleans
As per BSON specification, a boolean value is a single byte, whereas an integer is at least a 32-bit integer consuming 4 bytes. So it does make a difference and if you have millions of entries that would amount to... ...well, a few megabytes. ;) So for all practical purposes there is no difference.