count

Command Fields

count has the following fields:

readConcern: { level: <value> }

collation: {
   locale: <string>,
   caseLevel: <boolean>,
   caseFirst: <string>,
   strength: <int>,
   numericOrdering: <boolean>,
   alternate: <string>,
   maxVariable: <string>,
   backwards: <boolean>
}

Inaccurate Counts Without Query Predicate

When you call count without a query predicate, you may receive inaccurate document counts. Without a query predicate, count commands return results based on the collection's metadata, which may result in an approximate count. In particular,

• On a sharded cluster, the resulting count will not correctly filter out orphaned documents.

• After an unclean shutdown or file copy based initial sync, the count may be incorrect.

For counts based on collection metadata, see also collStats pipeline stage with the count option.

Count and Transactions

When you use count in a transaction, the resulting count will not filter out any uncommitted multi-document transactions.

For details, see Transactions and Count Operations.

Accuracy and Sharded Clusters

On a sharded cluster, the count command when run without a query predicate can result in an inaccurate count if orphaned documents exist or if a chunk migration is in progress.

To avoid these situations, on a sharded cluster, use the db.collection.aggregate() method:

You can use the $count stage to count the documents. For example, the following operation counts the documents in a collection:

db.collection.aggregate( [
   { $count: "myCount" }
])

The $count stage is equivalent to the following $group + $project sequence:

db.collection.aggregate( [
   { $group: { _id: null, count: { $sum: 1 } } }
   { $project: { _id: 0 } }
] )

Accuracy after Unexpected Shutdown

After an unclean shutdown of a mongod using the Wired Tiger storage engine, count statistics reported by count may be inaccurate.

The amount of drift depends on the number of insert, update, or delete operations performed between the last checkpoint and the unclean shutdown. Checkpoints usually occur every 60 seconds. However, mongod instances running with non-default --syncdelay settings may have more or less frequent checkpoints.

Run validate on each collection on the mongod to restore statistics after an unclean shutdown.

After an unclean shutdown:

• validate updates the count statistic in the collStats output with the latest value.

• Other statistics like the number of documents inserted or removed in the collStats output are estimates.

Client Disconnection

Starting in MongoDB 4.2, if the client that issued count disconnects before the operation completes, MongoDB marks count for termination using killOp.

Count All Documents

The following operation counts the number of all documents in the orders collection:

db.runCommand( { count: 'orders' } )

In the result, the n, which represents the count, is 26, and the command status ok is 1:

{ "n" : 26, "ok" : 1 }

Count Documents That Match a Query

The following operation returns a count of the documents in the orders collection where the value of the ord_dt field is greater than Date('01/01/2012'):

db.runCommand( { count:'orders',
                 query: { ord_dt: { $gt: new Date('01/01/2012') } }
               } )

In the result, the n, which represents the count, is 13 and the command status ok is 1:

{ "n" : 13, "ok" : 1 }

Skip Documents in Count

The following operation returns a count of the documents in the orders collection where the value of the ord_dt field is greater than Date('01/01/2012') and skip the first 10 matching documents:

db.runCommand( { count:'orders',
                 query: { ord_dt: { $gt: new Date('01/01/2012') } },
                 skip: 10 }  )

In the result, the n, which represents the count, is 3 and the command status ok is 1:

{ "n" : 3, "ok" : 1 }

Specify the Index to Use

The following operation uses the index { status: 1 } to return a count of the documents in the orders collection where the value of the ord_dt field is greater than Date('01/01/2012') and the status field is equal to "D":

db.runCommand(
   {
     count:'orders',
     query: {
              ord_dt: { $gt: new Date('01/01/2012') },
              status: "D"
            },
     hint: { status: 1 }
   }
)

In the result, the n, which represents the count, is 1 and the command status ok is 1:

{ "n" : 1, "ok" : 1 }

Override Default Read Concern

To override the default read concern level of "local", use the readConcern option.

The following operation on a replica set specifies a Read Concern of "majority" to read the most recent copy of the data confirmed as having been written to a majority of the nodes.

db.runCommand(
   {
     count: "restaurants",
     query: { rating: { $gte: 4 } },
     readConcern: { level: "majority" }
   }
)

To ensure that a single thread can read its own writes, use "majority" read concern and "majority" write concern against the primary of the replica set.