Apache Druid
  • Technology
  • Use Cases
  • Powered By
  • Docs
  • Community
  • Apache
  • Download

›Tutorials

Getting started

  • Introduction to Apache Druid
  • Quickstart (local)
  • Single server deployment
  • Clustered deployment

Tutorials

  • Load files natively
  • Load files using SQL 🆕
  • Load from Apache Kafka
  • Load from Apache Hadoop
  • Querying data
  • Roll-up
  • Theta sketches
  • Configuring data retention
  • Updating existing data
  • Compacting segments
  • Deleting data
  • Writing an ingestion spec
  • Transforming input data
  • Tutorial: Run with Docker
  • Kerberized HDFS deep storage
  • Convert ingestion spec to SQL
  • Jupyter Notebook tutorials

Design

  • Design
  • Segments
  • Processes and servers
  • Deep storage
  • Metadata storage
  • ZooKeeper

Ingestion

  • Ingestion
  • Data formats
  • Data model
  • Data rollup
  • Partitioning
  • Ingestion spec
  • Schema design tips
  • Stream ingestion

    • Apache Kafka ingestion
    • Apache Kafka supervisor
    • Apache Kafka operations
    • Amazon Kinesis

    Batch ingestion

    • Native batch
    • Native batch: input sources
    • Migrate from firehose
    • Hadoop-based

    SQL-based ingestion 🆕

    • Overview
    • Key concepts
    • API
    • Security
    • Examples
    • Reference
    • Known issues
  • Task reference
  • Troubleshooting FAQ

Data management

  • Overview
  • Data updates
  • Data deletion
  • Schema changes
  • Compaction
  • Automatic compaction

Querying

    Druid SQL

    • Overview and syntax
    • SQL data types
    • Operators
    • Scalar functions
    • Aggregation functions
    • Multi-value string functions
    • JSON functions
    • All functions
    • Druid SQL API
    • JDBC driver API
    • SQL query context
    • SQL metadata tables
    • SQL query translation
  • Native queries
  • Query execution
  • Troubleshooting
  • Concepts

    • Datasources
    • Joins
    • Lookups
    • Multi-value dimensions
    • Nested columns
    • Multitenancy
    • Query caching
    • Using query caching
    • Query context

    Native query types

    • Timeseries
    • TopN
    • GroupBy
    • Scan
    • Search
    • TimeBoundary
    • SegmentMetadata
    • DatasourceMetadata

    Native query components

    • Filters
    • Granularities
    • Dimensions
    • Aggregations
    • Post-aggregations
    • Expressions
    • Having filters (groupBy)
    • Sorting and limiting (groupBy)
    • Sorting (topN)
    • String comparators
    • Virtual columns
    • Spatial filters

Configuration

  • Configuration reference
  • Extensions
  • Logging

Operations

  • Web console
  • Java runtime
  • Security

    • Security overview
    • User authentication and authorization
    • LDAP auth
    • Password providers
    • Dynamic Config Providers
    • TLS support

    Performance tuning

    • Basic cluster tuning
    • Segment size optimization
    • Mixed workloads
    • HTTP compression
    • Automated metadata cleanup

    Monitoring

    • Request logging
    • Metrics
    • Alerts
  • API reference
  • High availability
  • Rolling updates
  • Using rules to drop and retain data
  • Working with different versions of Apache Hadoop
  • Misc

    • dump-segment tool
    • reset-cluster tool
    • insert-segment-to-db tool
    • pull-deps tool
    • Deep storage migration
    • Export Metadata Tool
    • Metadata Migration
    • Content for build.sbt

Development

  • Developing on Druid
  • Creating extensions
  • JavaScript functionality
  • Build from source
  • Versioning
  • Experimental features

Misc

  • Papers

Hidden

  • Apache Druid vs Elasticsearch
  • Apache Druid vs. Key/Value Stores (HBase/Cassandra/OpenTSDB)
  • Apache Druid vs Kudu
  • Apache Druid vs Redshift
  • Apache Druid vs Spark
  • Apache Druid vs SQL-on-Hadoop
  • Authentication and Authorization
  • Broker
  • Coordinator Process
  • Historical Process
  • Indexer Process
  • Indexing Service
  • MiddleManager Process
  • Overlord Process
  • Router Process
  • Peons
  • Approximate Histogram aggregators
  • Apache Avro
  • Microsoft Azure
  • Bloom Filter
  • DataSketches extension
  • DataSketches HLL Sketch module
  • DataSketches Quantiles Sketch module
  • DataSketches Theta Sketch module
  • DataSketches Tuple Sketch module
  • Basic Security
  • Kerberos
  • Cached Lookup Module
  • Apache Ranger Security
  • Google Cloud Storage
  • HDFS
  • Apache Kafka Lookups
  • Globally Cached Lookups
  • MySQL Metadata Store
  • ORC Extension
  • Druid pac4j based Security extension
  • Apache Parquet Extension
  • PostgreSQL Metadata Store
  • Protobuf
  • S3-compatible
  • Simple SSLContext Provider Module
  • Stats aggregator
  • Test Stats Aggregators
  • Druid AWS RDS Module
  • Kubernetes
  • Ambari Metrics Emitter
  • Apache Cassandra
  • Rackspace Cloud Files
  • DistinctCount Aggregator
  • Graphite Emitter
  • InfluxDB Line Protocol Parser
  • InfluxDB Emitter
  • Kafka Emitter
  • Materialized View
  • Moment Sketches for Approximate Quantiles module
  • Moving Average Query
  • OpenTSDB Emitter
  • Druid Redis Cache
  • Microsoft SQLServer
  • StatsD Emitter
  • T-Digest Quantiles Sketch module
  • Thrift
  • Timestamp Min/Max aggregators
  • GCE Extensions
  • Aliyun OSS
  • Prometheus Emitter
  • kubernetes
  • Cardinality/HyperUnique aggregators
  • Select
  • Firehose (deprecated)
  • Native batch (simple)
  • Realtime Process
Edit

Approximations with Theta sketches

Apache Druid can power real-time collection, streaming, and interactive visualization of clickstreams. A common problem in clickstream analytics is counting unique things, like visitors or sessions. Generally this involves scanning through all detail data, because unique counts do not add up as you aggregate the numbers.

The problem with counts and set operations on large data sets

Imagine you are interested in the number of visitors that watched episodes of a TV show. Let's say you found that at a given day, 1000 unique visitors watched the first episode, and 800 visitors watched the second episode. You may want to explore further trends, for example:

  • How many visitors watched both episodes?
  • How many visitors are there that watched at least one of the episodes?
  • How many visitors watched episode 1 but not episode 2?

There is no way to answer these questions by just looking at the aggregated numbers. You would have to go back to the detail data and scan every single row. If the data volume is high enough, this may take a very long time, meaning that an interactive data exploration is not possible.

An additional nuisance is that unique counts don't work well with rollups. For this example, it would be great if you could have just one row of data per 15 minute interval[1], show, and episode. After all, you are not interested in the individual user IDs, just the unique counts.

Is there a way to avoid crunching the detail data every single time, and maybe even enable rollup? Enter Theta sketches.

Use Theta sketches for fast approximation with set operations

Use Theta sketches to obtain a fast approximate estimate for the distinct count of values used to build the sketches. Theta sketches are a probabilistic data structure to enable approximate analysis of big data with known error distributions. Druid's implementation relies on the Apache DataSketches library.

The following properties describe Theta sketches:

  • Similar to other sketches, Theta sketches are mergeable. This means you can work with rolled up data and merge the sketches over various time intervals. Thus, you can take advantage of Druid's rollup feature.
  • Specific to sketches supported in Druid, Theta sketches support set operations. Given two Theta sketches over subsets of data, you can compute the union, intersection, or set difference of the two subsets. This enables you to answer questions like the number of visitors that watched a specific combination of episodes from the example.

In this tutorial, you will learn how to do the following:

  • Create Theta sketches from your input data at ingestion time.
  • Execute distinct count and set operation queries on the Theta sketches to explore the questions presented earlier.

Prerequisites

For this tutorial, you should have already downloaded Druid as described in the single-machine quickstart and have it running on your local machine. It will also be helpful to have finished Tutorial: Loading a file and Tutorial: Querying data.

This tutorial works with the following data:

  • date: a timestamp. In this case it's just dates but as mentioned earlier, a finer granularity makes sense in real life.
  • uid: a user ID
  • show: name of a TV show
  • episode: episode identifier
date,uid,show,episode
2022-05-19,alice,Game of Thrones,S1E1
2022-05-19,alice,Game of Thrones,S1E2
2022-05-19,alice,Game of Thrones,S1E1
2022-05-19,bob,Bridgerton,S1E1
2022-05-20,alice,Game of Thrones,S1E1
2022-05-20,carol,Bridgerton,S1E2
2022-05-20,dan,Bridgerton,S1E1
2022-05-21,alice,Game of Thrones,S1E1
2022-05-21,carol,Bridgerton,S1E1
2022-05-21,erin,Game of Thrones,S1E1
2022-05-21,alice,Bridgerton,S1E1
2022-05-22,bob,Game of Thrones,S1E1
2022-05-22,bob,Bridgerton,S1E1
2022-05-22,carol,Bridgerton,S1E2
2022-05-22,bob,Bridgerton,S1E1
2022-05-22,erin,Game of Thrones,S1E1
2022-05-22,erin,Bridgerton,S1E2
2022-05-23,erin,Game of Thrones,S1E1
2022-05-23,alice,Game of Thrones,S1E1

Ingest data using Theta sketches

  1. Navigate to the Load data wizard in the web console.
  2. Select Paste data as the data source and paste the given data:

Load data view with pasted data

  1. Leave the source type as inline and click Apply and Next: Parse data.
  2. Parse the data as CSV, with included headers:

Parse raw data

  1. Accept the default values in the Parse time, Transform, and Filter stages.
  2. In the Configure schema stage, enable rollup and confirm your choice in the dialog. Then set the query granularity to day.

Configure schema for rollup and query granularity

  1. Add the Theta sketch during this stage. Select Add metric.
  2. Define the new metric as a Theta sketch with the following details:
    • Name: theta_uid
    • Type: thetaSketch
    • Field name: uid
    • Size: Accept the default value, 16384.
    • Is input theta sketch: Accept the default value, False.

Create Theta sketch metric

  1. Click Apply to add the new metric to the data model.
  1. You are not interested in individual user ID's, only the unique counts. Right now, uid is still in the data model. To remove it, click on the uid column in the data model and delete it using the trashcan icon on the right:

Delete uid column

  1. For the remaining stages of the Load data wizard, set the following options:
    • Partition: Set Segment granularity to day.
    • Tune: Leave the default options.
    • Publish: Set the datasource name to ts_tutorial.

On the Edit spec page, your final input spec should match the following:

{
  "type": "index_parallel",
  "spec": {
    "ioConfig": {
      "type": "index_parallel",
      "inputSource": {
        "type": "inline",
        "data": "date,uid,show,episode\n2022-05-19,alice,Game of Thrones,S1E1\n2022-05-19,alice,Game of Thrones,S1E2\n2022-05-19,alice,Game of Thrones,S1E1\n2022-05-19,bob,Bridgerton,S1E1\n2022-05-20,alice,Game of Thrones,S1E1\n2022-05-20,carol,Bridgerton,S1E2\n2022-05-20,dan,Bridgerton,S1E1\n2022-05-21,alice,Game of Thrones,S1E1\n2022-05-21,carol,Bridgerton,S1E1\n2022-05-21,erin,Game of Thrones,S1E1\n2022-05-21,alice,Bridgerton,S1E1\n2022-05-22,bob,Game of Thrones,S1E1\n2022-05-22,bob,Bridgerton,S1E1\n2022-05-22,carol,Bridgerton,S1E2\n2022-05-22,bob,Bridgerton,S1E1\n2022-05-22,erin,Game of Thrones,S1E1\n2022-05-22,erin,Bridgerton,S1E2\n2022-05-23,erin,Game of Thrones,S1E1\n2022-05-23,alice,Game of Thrones,S1E1"
      },
      "inputFormat": {
        "type": "csv",
        "findColumnsFromHeader": true
      }
    },
    "tuningConfig": {
      "type": "index_parallel",
      "partitionsSpec": {
        "type": "hashed"
      },
      "forceGuaranteedRollup": true
    },
    "dataSchema": {
      "dataSource": "ts_tutorial",
      "timestampSpec": {
        "column": "date",
        "format": "auto"
      },
      "dimensionsSpec": {
        "dimensions": [
          "show",
          "episode"
        ]
      },
      "granularitySpec": {
        "queryGranularity": "day",
        "rollup": true,
        "segmentGranularity": "day"
      },
      "metricsSpec": [
        {
          "name": "count",
          "type": "count"
        },
        {
          "type": "thetaSketch",
          "name": "theta_uid",
          "fieldName": "uid"
        }
      ]
    }
  }
}

Notice the theta_uid object in the metricsSpec list, that defines the thetaSketch aggregator on the uid column during ingestion.

Click Submit to start the ingestion.

Query the Theta sketch column

Calculating a unique count estimate from a Theta sketch column involves the following steps:

  1. Merge the Theta sketches in the column by means of the DS_THETA aggregator function in Druid SQL.
  2. Retrieve the estimate from the merged sketch with the THETA_SKETCH_ESTIMATE function.

Between steps 1 and 2, you can apply set functions as demonstrated later in Set operations.

Basic counting

Let's first see what the data looks like in Druid. Run the following SQL statement in the query editor:

SELECT * FROM ts_tutorial

View data with SELECT all query

The Theta sketch column theta_uid appears as a Base64-encoded string; behind it is a bitmap.

The following query to compute the distinct counts of user IDs uses APPROX_COUNT_DISTINCT_DS_THETA and groups by the other dimensions:

SELECT __time,
       "show",
       "episode",
       APPROX_COUNT_DISTINCT_DS_THETA(theta_uid) AS users
FROM   ts_tutorial
GROUP  BY 1, 2, 3

Count distinct with Theta sketches

In the preceding query, APPROX_COUNT_DISTINCT_DS_THETA is equivalent to calling DS_THETA and THETA_SKETCH_ESIMATE as follows:

SELECT __time,
       "show", 
       "episode",
       THETA_SKETCH_ESTIMATE(DS_THETA(theta_uid)) AS users
FROM   ts_tutorial
GROUP  BY 1, 2, 3

That is, APPROX_COUNT_DISTINCT_DS_THETA applies the following:

  • DS_THETA: Creates a new Theta sketch from the column of Theta sketches
  • THETA_SKETCH_ESTIMATE: Calculates the distinct count estimate from the output of DS_THETA

Filtered metrics

Druid has the capability to use filtered metrics. This means you can include a WHERE clause in the SELECT part of the query.

In the case of Theta sketches, the filter clause has to be inserted between the aggregator and the estimator.

As an example, query the total unique users that watched Bridgerton:

SELECT THETA_SKETCH_ESTIMATE(
         DS_THETA(theta_uid) FILTER(WHERE "show" = 'Bridgerton')
       ) AS users
FROM ts_tutorial

Count distinct with Theta sketches and filters

Set operations

You can use this capability of filtering in the aggregator, together with set operations, to finally answer the questions from the introduction.

How many users watched both episodes of Bridgerton? Use THETA_SKETCH_INTERSECT to compute the unique count of the intersection of two (or more) segments:

SELECT THETA_SKETCH_ESTIMATE(
         THETA_SKETCH_INTERSECT(
           DS_THETA(theta_uid) FILTER(WHERE "show" = 'Bridgerton' AND "episode" = 'S1E1'),
           DS_THETA(theta_uid) FILTER(WHERE "show" = 'Bridgerton' AND "episode" = 'S1E2')
         )
       ) AS users
FROM ts_tutorial

Count distinct with Theta sketches, filters, and set operations

Again, the set function is spliced in between the aggregator and the estimator.

Likewise, use THETA_SKETCH_UNION to find the number of visitors that watched any of the episodes:

SELECT THETA_SKETCH_ESTIMATE(
         THETA_SKETCH_UNION(
           DS_THETA(theta_uid) FILTER(WHERE "show" = 'Bridgerton' AND "episode" = 'S1E1'),
           DS_THETA(theta_uid) FILTER(WHERE "show" = 'Bridgerton' AND "episode" = 'S1E2')
         )
       ) AS users
FROM ts_tutorial

Count distinct with Theta sketches, filters, and set operations

And finally, there is THETA_SKETCH_NOT which computes the set difference of two or more segments. The result describes how many visitors watched episode 1 of Bridgerton but not episode 2.

SELECT THETA_SKETCH_ESTIMATE(
         THETA_SKETCH_NOT(
           DS_THETA(theta_uid) FILTER(WHERE "show" = 'Bridgerton' AND "episode" = 'S1E1'),
           DS_THETA(theta_uid) FILTER(WHERE "show" = 'Bridgerton' AND "episode" = 'S1E2')
         )
       ) AS users
FROM ts_tutorial

Count distinct with Theta sketches, filters, and set operations

Conclusions

  • Counting distinct things for large data sets can be done with Theta sketches in Apache Druid.
  • This allows us to use rollup and discard the individual values, just retaining statistical approximations in the sketches.
  • With Theta sketch set operations, affinity analysis is easier, for example, to answer questions such as which segments correlate or overlap by how much.

Further reading

See the following topics for more information:

  • Theta sketch for reference on ingestion and native queries on Theta sketches in Druid.
  • Theta sketch scalar functions and Theta sketch aggregation functions for Theta sketch functions in Druid SQL queries.
  • Sketches for high cardinality columns for Druid schema design involving sketches.
  • DataSketches extension for more information about the DataSketches extension in Druid as well as other available sketches.
  • The accuracy of queries using Theta sketches is governed by the size k of the Theta sketch and by the operations you perform. See more details in the Apache DataSketches documentation.

Acknowledgments

This tutorial is adapted from a blog post by community member Hellmar Becker.


  1. Why 15 minutes and not just 1 hour? Intervals of 15 minutes work better with international timezones because those are not always aligned by hour. India, for instance, is 30 minutes off, and Nepal is even 45 minutes off. With 15 minute aggregates, you can get hourly sums for any of those timezones, too! ↩

← Roll-upConfiguring data retention →
  • The problem with counts and set operations on large data sets
  • Use Theta sketches for fast approximation with set operations
  • Prerequisites
  • Ingest data using Theta sketches
  • Query the Theta sketch column
    • Basic counting
    • Filtered metrics
    • Set operations
  • Conclusions
  • Further reading
  • Acknowledgments

Technology · Use Cases · Powered by Druid · Docs · Community · Download · FAQ

 ·  ·  · 
Copyright © 2022 Apache Software Foundation.
Except where otherwise noted, licensed under CC BY-SA 4.0.
Apache Druid, Druid, and the Druid logo are either registered trademarks or trademarks of The Apache Software Foundation in the United States and other countries.