union

The union operator in APL allows you to combine the results of two or more queries into a single output. The operator is useful when you need to analyze or compare data from different datasets or tables in a unified manner. By using union, you can merge multiple sets of records, keeping all data from the source tables without applying any aggregation or filtering.

The union operator is particularly helpful in scenarios like log analysis, tracing OpenTelemetry events, or correlating security logs across multiple sources. You can use it to perform comprehensive investigations by bringing together information from different datasets into one query.

​

Union of two datasets

To understand how the union operator works, consider these datasets:

Server requests

App logs

Performing a union on Server requests and Application logs would result in a new dataset with all the rows from both DatasetA and DatasetB.

A union of requests and logs would produce the following result set:

This result combines the rows and merges types for overlapping fields.

​

For users of other query languages

If you come from other query languages, this section explains how to adjust your existing queries to achieve the same results in APL.

​

Usage

​

Syntax

T1 | union [T2], [T3], ...

​

Parameters

• T1, T2, T3, ...: Tables or query results you want to combine into a single output.

​

Returns

The union operator returns all rows from the specified tables or queries. If fields overlap, they are merged. Non-overlapping fields are retained in their original form.

​

Use case examples

• Log analysis
• OpenTelemetry traces
• Security logs

['sample-http-logs']
| union ['security-logs']
| where status == '500'

​

Other examples

​

Basic union

This example combines all rows from github-push-event and github-pull-request-event without any transformation or filtering.

['github-push-event']
| union ['github-pull-request-event']

Run in Playground

​

Filter after union

This example combines the datasets, and then filters the data to only include rows where the method is GET.

['sample-http-logs']
| union ['github-issues-event']
| where method == "GET"

Run in Playground

​

Aggregate after union

This example combines the datasets and summarizes the data, counting the occurrences of each combination of content_type and actor.

['sample-http-logs']
| union ['github-pull-request-event']
| summarize Count = count() by content_type, actor

Run in Playground

​

Filter and project specific data from combined log sources

This query combines GitHub pull request event logs and GitHub push events, filters by actions made by github-actions[bot], and displays key event details such as time, repository, commits, head , id.

['github-pull-request-event']
| union ['github-push-event']
| where actor == "github-actions[bot]"
| project _time, repo, ['id'], commits, head

Run in Playground

​

Union with field removing

This example removes the content_type and commits field in the datasets sample-http-logs and github-push-event before combining the datasets.

['sample-http-logs']
| union ['github-push-event']
| project-away content_type, commits

Run in Playground

​

Filter after union

This example performs a union and then filters the resulting set to only include rows where the method is GET.

['sample-http-logs']
| union ['github-issues-event']
| where method == "GET"

Run in Playground

​

Union with order by

After the union, the result is ordered by the type field.

['sample-http-logs']
| union hn
| order by type

Run in Playground

​

Union with joint conditions

This example performs a union and then filters the resulting dataset for rows where content_type contains the letter a and city is seattle.

['sample-http-logs']
| union ['github-pull-request-event']
| where content_type contains "a" and ['geo.city']  == "Seattle"

Run in Playground

​

Union and count unique values

After the union, the query calculates the number of unique geo.city and repo entries in the combined dataset.

['sample-http-logs']
| union ['github-push-event']
| summarize UniqueNames = dcount(['geo.city']), UniqueData = dcount(repo)

Run in Playground

​

Best practices for the union operator

To maximize the effectiveness of the union operator in APL, here are some best practices to consider:

• Before using the union operator, ensure that the fields being merged have compatible data types.
• Use project or project-away to include or exclude specific fields. This can improve performance and the clarity of your results, especially when you only need a subset of the available data.