fork operator

Runs multiple consumer operators in parallel.

Syntax

T | fork [name=](subquery) [name=](subquery) …

Parameters

Supported query operators

• as
• count
• extend
• parse
• where
• take
• project
• project-away
• project-keep
• project-rename
• project-reorder
• summarize
• top
• top-nested
• sort
• mv-expand
• reduce

Returns

Multiple result tables, one for each of the subquery arguments.

Tips

• Use materialize as a replacement for join or union on fork legs. The input stream is cached by materialize and then the cached expression can be used in join/union legs.

• Use batch with materialize of tabular expression statements instead of the fork operator.

Examples

The examples output multiple tables, with named and umnamed columns.

Unnamed subqueries

StormEvents
| where State == "FLORIDA"
| fork
    ( where DeathsDirect + DeathsIndirect > 1)
    ( where InjuriesDirect + InjuriesIndirect > 1)

Output

This output shows the first few rows and columns of the result table.

GenericResult

GenericResult

Named subqueries

In the following examples, the result table is named “StormsWithDeaths” and “StormsWithInjuries”.

StormEvents
| where State == "FLORIDA"
| fork
    (where DeathsDirect + DeathsIndirect > 1 | as StormsWithDeaths)
    (where InjuriesDirect + InjuriesIndirect > 1 | as StormsWithInjuries)

StormEvents
| where State == "FLORIDA"
| fork
    StormsWithDeaths = (where DeathsDirect + DeathsIndirect > 1)
    StormsWithInjuries = (where InjuriesDirect + InjuriesIndirect > 1)

Output

This output shows the first few rows and columns of the result table.

StormsWithDeaths

StormsWithInjuries

SamplePowerRequirementHistorizedData
| fork
    Dataset2 = (where twinId  <> "p_sol_01" | summarize count() by twinId, name)
    Dataset3 = (summarize count() by WeekOfYear = week_of_year(timestamp))

It is possible to use almost all the known features of the KQL language inside every single “sub” result set. For instance, the join operator inside a sub-statement does not work. This is not allowed by the engine.