LogoLogo
release-1.3.0
release-1.3.0
  • Introduction
  • Basics
    • Concepts
      • Pinot storage model
      • Architecture
      • Components
        • Cluster
          • Tenant
          • Server
          • Controller
          • Broker
          • Minion
        • Table
          • Segment
            • Deep Store
            • Segment threshold
            • Segment retention
          • Schema
          • Time boundary
        • Pinot Data Explorer
    • Getting Started
      • Running Pinot locally
      • Running Pinot in Docker
      • Quick Start Examples
      • Running in Kubernetes
      • Running on public clouds
        • Running on Azure
        • Running on GCP
        • Running on AWS
      • Create and update a table configuration
      • Batch import example
      • Stream ingestion example
      • HDFS as Deep Storage
      • Troubleshooting Pinot
      • Frequently Asked Questions (FAQs)
        • General
        • Pinot On Kubernetes FAQ
        • Ingestion FAQ
        • Query FAQ
        • Operations FAQ
    • Import Data
      • From Query Console
      • Batch Ingestion
        • Spark
        • Flink
        • Hadoop
        • Backfill Data
        • Dimension table
      • Stream ingestion
        • Ingest streaming data from Apache Kafka
        • Ingest streaming data from Amazon Kinesis
        • Ingest streaming data from Apache Pulsar
        • Configure indexes
      • Stream ingestion with Upsert
      • Segment compaction on upserts
      • Stream ingestion with Dedup
      • Stream ingestion with CLP
      • File Systems
        • Amazon S3
        • Azure Data Lake Storage
        • HDFS
        • Google Cloud Storage
      • Input formats
        • Complex Type (Array, Map) Handling
        • Complex Type Examples
        • Ingest records with dynamic schemas
      • Reload a table segment
      • Upload a table segment
    • Indexing
      • Bloom filter
      • Dictionary index
      • Forward index
      • FST index
      • Geospatial
      • Inverted index
      • JSON index
      • Native text index
      • Range index
      • Star-tree index
      • Text search support
      • Timestamp index
      • Vector index
    • Release notes
      • 1.3.0
      • 1.2.0
      • 1.1.0
      • 1.0.0
      • 0.12.1
      • 0.12.0
      • 0.11.0
      • 0.10.0
      • 0.9.3
      • 0.9.2
      • 0.9.1
      • 0.9.0
      • 0.8.0
      • 0.7.1
      • 0.6.0
      • 0.5.0
      • 0.4.0
      • 0.3.0
      • 0.2.0
      • 0.1.0
    • Recipes
      • Connect to Streamlit
      • Connect to Dash
      • Visualize data with Redash
      • GitHub Events Stream
  • For Users
    • Query
      • Querying Pinot
      • Query Syntax
        • Aggregation Functions
        • Array Functions
        • Cardinality Estimation
        • Explain Plan (Single-Stage)
        • Filtering with IdSet
        • Funnel Analysis
        • GapFill Function For Time-Series Dataset
        • Grouping Algorithm
        • Hash Functions
        • JOINs
        • Lookup UDF Join
        • Querying JSON data
        • Transformation Functions
        • URL Functions
        • Window Functions
      • Query Options
      • Query Quotas
      • Query using Cursors
      • Multi-stage query
        • Understanding Stages
        • Stats
        • Optimizing joins
        • Join strategies
          • Random + broadcast join strategy
          • Query time partition join strategy
          • Colocated join strategy
          • Lookup join strategy
        • Hints
        • Operator Types
          • Aggregate
          • Filter
          • Join
          • Intersect
          • Leaf
          • Literal
          • Mailbox receive
          • Mailbox send
          • Minus
          • Sort or limit
          • Transform
          • Union
          • Window
        • Stage-Level Spooling
      • User-Defined Functions (UDFs)
      • Explain plan
    • APIs
      • Broker Query API
        • Query Response Format
      • Controller Admin API
      • Controller API Reference
    • External Clients
      • JDBC
      • Java
      • Python
      • Golang
    • Tutorials
      • Use OSS as Deep Storage for Pinot
      • Ingest Parquet Files from S3 Using Spark
      • Creating Pinot Segments
      • Use S3 as Deep Storage for Pinot
      • Use S3 and Pinot in Docker
      • Batch Data Ingestion In Practice
      • Schema Evolution
  • For Developers
    • Basics
      • Extending Pinot
        • Writing Custom Aggregation Function
        • Segment Fetchers
      • Contribution Guidelines
      • Code Setup
      • Code Modules and Organization
      • Dependency Management
      • Update documentation
    • Advanced
      • Data Ingestion Overview
      • Ingestion Aggregations
      • Ingestion Transformations
      • Null value support
      • Use the multi-stage query engine (v2)
      • Advanced Pinot Setup
    • Plugins
      • Write Custom Plugins
        • Input Format Plugin
        • Filesystem Plugin
        • Batch Segment Fetcher Plugin
        • Stream Ingestion Plugin
    • Design Documents
      • Segment Writer API
  • For Operators
    • Deployment and Monitoring
      • Set up cluster
      • Server Startup Status Checkers
      • Set up table
      • Set up ingestion
      • Decoupling Controller from the Data Path
      • Segment Assignment
      • Instance Assignment
      • Rebalance
        • Rebalance Servers
        • Rebalance Brokers
        • Rebalance Tenant
      • Separating data storage by age
        • Using multiple tenants
        • Using multiple directories
      • Pinot managed Offline flows
      • Minion merge rollup task
      • Consistent Push and Rollback
      • Access Control
      • Monitoring
      • Tuning
        • Tuning Default MMAP Advice
        • Real-time
        • Routing
        • Query Routing using Adaptive Server Selection
        • Query Scheduling
      • Upgrading Pinot with confidence
      • Managing Logs
      • OOM Protection Using Automatic Query Killing
      • Pause ingestion based on resource utilization
    • Command-Line Interface (CLI)
    • Configuration Recommendation Engine
    • Tutorials
      • Authentication
        • Basic auth access control
        • ZkBasicAuthAccessControl
      • Configuring TLS/SSL
      • Build Docker Images
      • Running Pinot in Production
      • Kubernetes Deployment
      • Amazon EKS (Kafka)
      • Amazon MSK (Kafka)
      • Monitor Pinot using Prometheus and Grafana
      • Performance Optimization Configurations
      • Segment Operations Throttling
  • Configuration Reference
    • Cluster
    • Controller
    • Broker
    • Server
    • Table
    • Ingestion
    • Schema
    • Ingestion Job Spec
    • Monitoring Metrics
    • Functions
      • ABS
      • ADD
      • ago
      • EXPR_MIN / EXPR_MAX
      • ARRAY_AGG
      • arrayConcatDouble
      • arrayConcatFloat
      • arrayConcatInt
      • arrayConcatLong
      • arrayConcatString
      • arrayContainsInt
      • arrayContainsString
      • arrayDistinctInt
      • arrayDistinctString
      • arrayIndexOfInt
      • arrayIndexOfString
      • ARRAYLENGTH
      • arrayRemoveInt
      • arrayRemoveString
      • arrayReverseInt
      • arrayReverseString
      • arraySliceInt
      • arraySliceString
      • arraySortInt
      • arraySortString
      • arrayUnionInt
      • arrayUnionString
      • AVGMV
      • Base64
      • caseWhen
      • ceil
      • CHR
      • codepoint
      • concat
      • count
      • COUNTMV
      • COVAR_POP
      • COVAR_SAMP
      • day
      • dayOfWeek
      • dayOfYear
      • DISTINCT
      • DISTINCTAVG
      • DISTINCTAVGMV
      • DISTINCTCOUNT
      • DISTINCTCOUNTBITMAP
      • DISTINCTCOUNTBITMAPMV
      • DISTINCTCOUNTHLL
      • DISTINCTCOUNTSMARTHLL
      • DISTINCTCOUNTHLLPLUS
      • DISTINCTCOUNTHLLMV
      • DISTINCTCOUNTMV
      • DISTINCTCOUNTRAWHLL
      • DISTINCTCOUNTRAWHLLMV
      • DISTINCTCOUNTRAWTHETASKETCH
      • DISTINCTCOUNTTHETASKETCH
      • DISTINCTCOUNTULL
      • DISTINCTSUM
      • DISTINCTSUMMV
      • DIV
      • DATETIMECONVERT
      • DATETRUNC
      • exp
      • FIRSTWITHTIME
      • FLOOR
      • FrequentLongsSketch
      • FrequentStringsSketch
      • FromDateTime
      • FromEpoch
      • FromEpochBucket
      • FUNNELCOUNT
      • FunnelCompleteCount
      • FunnelMaxStep
      • FunnelMatchStep
      • Histogram
      • hour
      • isSubnetOf
      • JSONFORMAT
      • JSONPATH
      • JSONPATHARRAY
      • JSONPATHARRAYDEFAULTEMPTY
      • JSONPATHDOUBLE
      • JSONPATHLONG
      • JSONPATHSTRING
      • jsonextractkey
      • jsonextractscalar
      • LAG
      • LASTWITHTIME
      • LEAD
      • length
      • ln
      • lower
      • lpad
      • ltrim
      • max
      • MAXMV
      • MD5
      • millisecond
      • min
      • minmaxrange
      • MINMAXRANGEMV
      • MINMV
      • minute
      • MOD
      • mode
      • month
      • mult
      • now
      • percentile
      • percentileest
      • percentileestmv
      • percentilemv
      • percentiletdigest
      • percentiletdigestmv
      • percentilekll
      • percentilerawkll
      • percentilekllmv
      • percentilerawkllmv
      • quarter
      • regexpExtract
      • regexpReplace
      • remove
      • replace
      • reverse
      • round
      • roundDecimal
      • ROW_NUMBER
      • rpad
      • rtrim
      • second
      • SEGMENTPARTITIONEDDISTINCTCOUNT
      • sha
      • sha256
      • sha512
      • sqrt
      • startswith
      • ST_AsBinary
      • ST_AsText
      • ST_Contains
      • ST_Distance
      • ST_GeogFromText
      • ST_GeogFromWKB
      • ST_GeometryType
      • ST_GeomFromText
      • ST_GeomFromWKB
      • STPOINT
      • ST_Polygon
      • strpos
      • ST_Union
      • SUB
      • substr
      • sum
      • summv
      • TIMECONVERT
      • timezoneHour
      • timezoneMinute
      • ToDateTime
      • ToEpoch
      • ToEpochBucket
      • ToEpochRounded
      • TOJSONMAPSTR
      • toGeometry
      • toSphericalGeography
      • trim
      • upper
      • Url
      • UTF8
      • VALUEIN
      • week
      • year
      • Extract
      • yearOfWeek
      • FIRST_VALUE
      • LAST_VALUE
      • ST_GeomFromGeoJSON
      • ST_GeogFromGeoJSON
      • ST_AsGeoJSON
    • Plugin Reference
      • Stream Ingestion Connectors
      • VAR_POP
      • VAR_SAMP
      • STDDEV_POP
      • STDDEV_SAMP
    • Dynamic Environment
  • Reference
    • Single-stage query engine (v1)
    • Multi-stage query engine (v2)
    • Troubleshooting
      • Troubleshoot issues with the multi-stage query engine (v2)
      • Troubleshoot issues with ZooKeeper znodes
  • RESOURCES
    • Community
    • Team
    • Blogs
    • Presentations
    • Videos
  • Integrations
    • Tableau
    • Trino
    • ThirdEye
    • Superset
    • Presto
    • Spark-Pinot Connector
  • Contributing
    • Contribute Pinot documentation
    • Style guide
Powered by GitBook
On this page
  • How to enable colocated joins
  • How to guarantee that colocated joins can be used
  • How to verify colocated joins are being used

Was this helpful?

Export as PDF
  1. For Users
  2. Query
  3. Multi-stage query
  4. Join strategies

Colocated join strategy

Colocated joins is a strategy Pinot can use execute joins without shuffling data between servers when all the following conditions are met:

  1. Both tables are partitioned.

  2. The partition function is the same.

  3. The number of partitions is the same.

  4. The join condition is an equality between the partition columns.

  5. The assignment of partitions to servers is the same.

  6. For each partition, there is a server that has all the segments of the partition for both tables.

As an example, imagine we want to execute the query

SELECT A.col1, B.col2 
FROM A
JOIN B
ON A.partitionKeyA = B.partitionKeyB

in a scenario where we have tables A and B partitioned by the same function in exactly two partitions and distributed in such a way that:

  1. The 1st partition of A is formed by segments A0-1 and A0-2, stored on servers 1 and 3.

  2. The 2nd partition of A is formed by segments A1-1 and A1-2, stored on server 2.

  3. The 1st partition of B is formed by segment B0-1, stored on server 3.

  4. The 2nd partition of B is formed by segment B1-1, stored on server 2.

In this case, Pinot will try to execute the query in the following way:

How to enable colocated joins

Colocated join optimization is disabled by default in Pinot 1.3.0.

It can be enabled cluster wise by setting the following configuration in the broker:

pinot.broker.multistage.infer.partition.hint=true

It can also be enabled/disabled on a per-query basis by setting the following query option:

SET inferPartitionHint=true
SELECT ...
Even lower level configuration (not recommended)

Colocated joins can also be enabled per-join basis by setting the tableOptions hint directly.

SELECT A.col1, B.col2
FROM A /*+ tableOptions(partition_function='hashcode', partition_key='partitionKeyA', partition_size='4') */
JOIN B /*+ tableOptions(partition_function='hashcode', partition_key='partitionKeyB', partition_size='4') */
ON A.partitionKeyA = B.partitionKeyB

In this case, the partition_function, partition_key, and partition_size are required to be the same for both tables and they must be the same as the ones defined in the table configuration.

This is a very advance and error prone way to configure joins that can also be used to change stage parallelism. Therefore it is recommended.

How to guarantee that colocated joins can be used

As noticed above, in order to use colocated joins, the assignment of partitions to servers must be the same for both tables. Although we can manually assign partitions to servers when creating the tables, they can be moved between servers at any time as a result of a rebalance.

How to verify colocated joins are being used

Notice that this optimization cannot be seen in the normal EXPLAIN PLAN command.

PreviousQuery time partition join strategyNextLookup join strategy

Was this helpful?

As a side effect, this strategy may not use as many servers as other techniques. For example, the same query using may use 3 servers, while in this case Pinot can only use server 3 and server 2. Server 1 cannot be used because it does not have all the segments for partition 2 of table B.

In order to guarantee that colocated joins can be used it is recommended to instruct Pinot to for each partition in both tables. To read more about how to partition a table, see and .

As explained, the main advantage when this optimization is enabled is that data doesn't need to be shuffled to execute the join. That can be verified with the and stats on the operator for this stage. All messages should be in memory and rawMessages should be 0 (or not listed at all).

Another way to verify this optimization is being applied is to use the command. You need to use the EXPLAIN IMPLEMENTATION PLAN command. There you will see that MAIL_SEND operators are decorated with [PARTITIONED] and each MAIL_SEND will send the data to another worker in the same server.

query time partition
assign the same instances
Instance Assignment
mailbox send
EXPLAIN IMPLEMENTATION PLAN
Routing
Dotted arrows mean shuffle while solid arrows mean in-server transfer
rawMessages
inMemoryMessages