prev()

Returns the value of a specific column in a specified row. The specified row is at a specified offset from the current row in a serialized row set.

Syntax

prev(column, [ offset ], [ default_value ] )

Parameters

Examples

Filter data based on comparison between adjacent rows

The following query returns rows that show breaks longer than a quarter of a second between calls to sensor-9.

TransformedSensorsData
| where SensorName == 'sensor-9'
| sort by Timestamp asc
| extend timeDiffInMilliseconds = datetime_diff('millisecond', Timestamp, prev(Timestamp, 1))
| where timeDiffInMilliseconds > 250

Output

Perform aggregation based on comparison between adjacent rows

The following query calculates the average time difference in milliseconds between calls to sensor-9.

TransformedSensorsData
| where SensorName == 'sensor-9'
| sort by Timestamp asc
| extend timeDiffInMilliseconds = datetime_diff('millisecond', Timestamp, prev(Timestamp, 1))
| summarize avg(timeDiffInMilliseconds)

Output

Extend row with data from the previous row

In the following query, as part of the serialization done with the serialize operator, a new column previous_session_type is added with data from the previous row. Since there was no session prior to the first session, the column is empty in the first row.

ConferenceSessions
| where conference == 'Build 2019'
| serialize previous_session_type = prev(session_type)
| project time_and_duration, session_title, session_type, previous_session_type

Output