Back to DocumentationMQL Documentation

$locf

Propagates the last observed non-null value forward within a partition, useful for filling missing data points in time-series datasets.

Syntax

{ $locf: { input: <expression>, sortBy: <document> } }

Parameters

inputobjectrequired

Expression that resolves to the field whose value to propagate

sortByobjectrequired

Document specifying the sort order for the partition

Examples

Fill missing time series data

Propagate last non-null value for missing data within company partition

Query:

db.stores.aggregate([{ $match: { "company": { $in: ["First Up Consultants"] } } }, { $setWindowFields: { "partitionBy": "$name", "sortBy": { "sales.revenue": 1 }, "output": { "lastUpdatedDate": { "$locf": { "lastUpdated": 1 } } } } }])

Output:

Stores with missing lastUpdated values filled with last observed non-null value

Forward fill sensor readings

Use last observed value to fill gaps in sensor data

Query:

db.sensors.aggregate([{ $setWindowFields: { "partitionBy": "$sensorId", "sortBy": { "timestamp": 1 }, "output": { "filledReading": { "$locf": { input: "$reading", sortBy: { "timestamp": 1 } } } } } }])

Output:

Sensor data with missing readings filled using last observed values

Related