robomongo | Native cross-platform MongoDB management tool

Is this query you want?

Using aggregate add all _ids to a set:

First of all, check if you are using latest Robo3T. MongoDB Atlas uses mongodb+srv:// protocol, that was not supported some time before.

After that please go by following check-list:

• Use import From SRV by copy-pasting URL from MongoDB Atlas;
• MongoDB Atlas is using Replica Sets even for free Sandbox (Atlas M0) accounts. So during creation ensure that you used that type;
• Fill or check all cluster members;
• Set Name should be set.

important differences on Authentication tab:

• Database should be admin not like database name (used on Heroku),
• Auth Mechanism should be SCRAM-SHA-1
• Check Manually specify visible databases and type there admin,

on TLS tab:

• Check Use TLS protocol
• Authentication method - Self-signed Certificate

Now it should work.

P.S. Previously, I also had problems with having database same like on Heroku with same db-name and user-name, also I thought that Auth Mechanism should be MONGODB-CR. But that didn't work.

Ok, it was an issue with Robo3T not with Atlas or permissions. For more details please refer to this feature update: Manually specify visible databases

In general, beginning from Robo 3T 1.4 you can and obviously you must manually specify visible database upon connection if your user hasn't admin privileges.

The motivation for this feature was this problem when the users without admin role privileges (or have access to a specific database), were unable to get the list of databases and so the connection was successful but there were no visible databases (as result of failed listdatabases command).

use ObjectId of mongo like this:

{ $match: { _id: mongo.ObjectId("5f43ebb18415f39cf1cd75d4") } }

complete example:

Nek's comment pointed me in the right direction, if you look at the docs for the mongoid.yml file, you will notice that there are two options sections, i had to move the authentication part inside one of them, the correct config looks like this:

You cannot connect to a replica set deployment through a tunnel, since the driver will (try to re) connect to the hostnames specified in replica set configuration as soon as it receives a response from any of the replica set members.

You can connect through a tunnel in single topology. Remove replicaSet URI option from your URI. Naturally this only gives you a connection to the specified node, you don't get automatic failover etc.

See also

I can see on front-end you have mentioned request type POST try changing it to DELETE

The problem is in TypedAggregation. This is a special aggregation where Spring holds information of the input aggregation type.

To avoid it, use raw aggregation (as if you run in MongoDB shell) this way:

[The following is a response to the original question, which was premised on JSON input.]

Since you need to count the items in a group, it would appear that you will find group_by( sub("/[^/]*$";"") ) useful.

For example, if you wanted to omit large groups entirely, as one interpretation of the stated requirements would seem to imply, you could use the following filter: