Stream ingestion with Upsert

Pinot provides native support of upserts during real-time ingestion. There are scenarios where records need modifications, such as correcting a ride fare or updating a delivery status.

Partial upserts are convenient as you only need to specify the columns where values change, and you ignore the rest.

Overview of upserts in Pinot

See an overview of how upserts work in Pinot 1.0.

Enable upserts in Pinot

To enable upserts on a Pinot table, do the following:

1. Define the primary key in the schema

2. Enable upserts in the table configurations

Define the primary key in the schema

To update a record, you need a primary key to uniquely identify the record. To define a primary key, add the field primaryKeyColumns to the schema definition. For example, the schema definition of UpsertMeetupRSVP in the quick start example has this definition.

{
    "primaryKeyColumns": ["event_id"]
}

Note this field expects a list of columns, as the primary key can be a composite.

When two records of the same primary key are ingested, the record with the greater comparison value (timeColumn by default) is used. When records have the same primary key and event time, then the order is not determined. In most cases, the later ingested record will be used, but this may not be true in cases where the table has a column to sort by.

Enable upsert in the table configurations

To enable upsert, make the following configurations in the table configurations.

Upsert modes

Full upsert

The upsert mode defaults to FULL . FULL upsert means that a new record will replace the older record completely if they have same primary key. Example config:

{
  "upsertConfig": {
    "mode": "FULL"
  }
}

Partial upserts

Partial upsert lets you choose to update only specific columns and ignore the rest.

To enable the partial upsert, set the mode to PARTIAL and specify partialUpsertStrategies for partial upsert columns. Since release-0.10.0, OVERWRITE is used as the default strategy for columns without a specified strategy. defaultPartialUpsertStrategy is also introduced to change the default strategy for all columns.

For example:

{
  "upsertConfig": {
    "mode": "PARTIAL",
    "partialUpsertStrategies":{
      "rsvp_count": "INCREMENT",
      "group_name": "IGNORE",
      "venue_name": "OVERWRITE"
    }
  },
  "tableIndexConfig": {
    "nullHandlingEnabled": true
  }
}

{
  "upsertConfig": {
    "mode": "PARTIAL",
    "defaultPartialUpsertStrategy": "OVERWRITE",
    "partialUpsertStrategies":{
      "rsvp_count": "INCREMENT",
      "group_name": "IGNORE"
    }
  },
  "tableIndexConfig": {
    "nullHandlingEnabled": true
  }
}

Pinot supports the following partial upsert strategies:

None upserts

If set mode to NONE, the upsert is disabled.

Comparison column

By default, Pinot uses the value in the time column (timeColumn in tableConfig) to determine the latest record. That means, for two records with the same primary key, the record with the larger value of the time column is picked as the latest update. However, there are cases when users need to use another column to determine the order. In such case, you can use option comparisonColumn to override the column used for comparison. For example,

{
  "upsertConfig": {
    "mode": "FULL",
    "comparisonColumn": "anotherTimeColumn"
  }
}

For partial upsert table, the out-of-order events won't be consumed and indexed. For example, for two records with the same primary key, if the record with the smaller value of the comparison column came later than the other record, it will be skipped.

Multiple comparison columns

In some cases, especially where partial upsert might be employed, there may be multiple producers of data each writing to a mutually exclusive set of columns, sharing only the primary key. In such a case, it may be helpful to use one comparison column per producer group so that each group can manage its own specific versioning semantics without the need to coordinate versioning across other producer groups.

{
  "upsertConfig": {
    "mode": "PARTIAL",
    "defaultPartialUpsertStrategy": "OVERWRITE",
    "partialUpsertStrategies":{},
    "comparisonColumns": ["secondsSinceEpoch", "otherComparisonColumn"]
  }
}

Documents written to Pinot are expected to have exactly 1 non-null value out of the set of comparisonColumns; if more than 1 of the columns contains a value, the document will be rejected. When new documents are written, whichever comparison column is non-null will be compared against only that same comparison column seen in prior documents with the same primary key. Consider the following examples, where the documents are assumed to arrive in the order specified in the array.

[
  {
    "event_id": "aa",
    "orderReceived": 1,
    "description" : "first",
    "secondsSinceEpoch": 1567205394
  },
  {
    "event_id": "aa",
    "orderReceived": 2,
    "description" : "update",
    "secondsSinceEpoch": 1567205397
  },
  {
    "event_id": "aa",
    "orderReceived": 3,
    "description" : "update",
    "secondsSinceEpoch": 1567205396
  },
  {
    "event_id": "aa",
    "orderReceived": 4,
    "description" : "first arrival, other column",
    "otherComparisonColumn": 1567205395
  },
  {
    "event_id": "aa",
    "orderReceived": 5,
    "description" : "late arrival, other column",
    "otherComparisonColumn": 1567205392
  },
  {
    "event_id": "aa",
    "orderReceived": 6,
    "description" : "update, other column",
    "otherComparisonColumn": 1567205398
  }
]

The following would occur:

1. orderReceived: 1

• Result: persisted

• Reason: first doc seen for primary key "aa"

1. orderReceived: 2

• Result: persisted (replacing orderReceived: 1)

• Reason: comparison column (secondsSinceEpoch) larger than that previously seen

1. orderReceived: 3

• Result: rejected

• Reason: comparison column (secondsSinceEpoch) smaller than that previously seen

1. orderReceived: 4

• Result: persisted (replacing orderReceived: 2)

• Reason: comparison column (otherComparisonColumn) larger than previously seen (never seen previously), despite the value being smaller than that seen for secondsSinceEpoch

1. orderReceived: 5

• Result: rejected

• Reason: comparison column (otherComparisonColumn) smaller than that previously seen

1. orderReceived: 6

• Result: persist (replacing orderReceived: 4)

• Reason: comparison column (otherComparisonColumn) larger than that previously seen

Metadata time-to-live (TTL)

In Pinot, the metadata map is stored in heap memory. To decrease in-memory data and improve performance, minimize the time primary key entries are stored in the metadata map (metadata time-to-live (TTL)). Limiting the TTL is especially useful for primary keys with high cardinality and frequent updates.

Since the metadata TTL is applied on the first comparison column, the time unit of upsert TTL is the same as the first comparison column.

Configure how long primary keys are stored in metadata

To configure how long primary keys are stored in metadata, specify the length of time in upsertTTL. For example:{

  "upsertConfig": {
    "mode": "FULL",
    "enableSnapshot": true,
    "enablePreload": true,
    "metadataTTL": 86400
  }
}

In this example, Pinot will retain primary keys in metadata for 1 day.

Note that enabling upsert snapshot is required for metadata TTL for in-memory validDocsIDs recovery.

Delete column

Upsert Pinot table can support soft-deletes of primary keys. This requires the incoming record to contain a dedicated boolean single-field column that serves as a delete marker for a primary key. Once the real-time engine encounters a record with delete column set to true , the primary key will no longer be part of the queryable set of documents. This means the primary key will not be visible in the queries, unless explicitly requested via query option skipUpsert=true.

{ 
    "upsertConfig": {  
        ... 
        "deleteRecordColumn": <column_name>
    } 
}

Note that the delete column has to be a single-value boolean column.

// In the Schema
{
    ...
    {
      "name": "<delete_column_name>",
      "dataType": "BOOLEAN"
    },
    ...
}

A deleted primary key can be revived by ingesting a record with the same primary, but with higher comparison column value(s).

Note that when reviving a primary key in a partial upsert table, the revived record will be treated as the source of truth for all columns. This means any previous updates to the columns will be ignored and overwritten with the new record's values.

Deleted Keys time-to-live (TTL)

The above config deleteRecordColumn only soft-deletes the primary key. To decrease in-memory data and improve performance, minimize the time deleted-primary-key entries are stored in the metadata map (deletedKeys time-to-live (TTL)). Limiting the TTL is especially useful for deleted-primary-keys where there are no future updates foreseen.

Configure how long deleted-primary-keys are stored in metadata

To configure how long primary keys are stored in metadata, specify the length of time in deletedKeysTTL For example:

  "upsertConfig": {
    "mode": "FULL",
    "deleteRecordColumn": <column_name>,
    "deletedKeysTTL": 86400
  }
}

In this example, Pinot will retain the deleted-primary-keys in metadata for 1 day.

Data consistency with deletes and compaction together

When using deletedKeysTTL together with UpsertCompactionTask, there can be a scenario where a segment containing deleted-record (where deleteRecordColumn = true was set for the primary key) gets compacted first and a previous old record is not yet compacted. During server restart, now the old record is added to the metadata manager map and is treated as non-deleted. To prevent data inconsistencies in this scenario, we have added a new config enableDeletedKeysCompactionConsistency which when set to true, will ensure that the deleted records are not compacted until all the previous records from all other segments are compacted for the deleted primary-key.

{
  "upsertConfig": {
    "mode": "FULL",
    "deleteRecordColumn": <column_name>,
    "deletedKeysTTL": 86400,
    "enableDeletedKeysCompactionConsistency": true
  }
}

Data consistency when queries and upserts happen concurrently

Upserts in Pinot enable real-time updates and ensure that queries always retrieve the latest version of a record, making them a powerful feature for managing mutable data efficiently. However, in applications with extremely high QPS and high ingestion rates, queries and upserts happening concurrently can sometimes lead to inconsistencies in query results.

For example, consider a table with 1 million primary keys. A distinct count query should always return 1 million, regardless of how new records are ingested and older records are invalidated. However, at high ingestion and query rates, the query may occasionally return a count slightly above or below 1 million. This happens because queries determine valid records by acquiring validDocIds bitmaps from multiple segments, which indicate which documents are currently valid. Since acquiring these bitmaps is not atomic with respect to ongoing upserts, a query may capture an inconsistent view of the data, leading to overcounting or undercounting of valid records.

This is a classic concurrency issue where reads and writes happen simultaneously, leading to temporary inconsistencies. Typically, such issues are resolved using locks or snapshots to maintain a stable view of the data during query execution. To address this, two new consistency modes - SYNC and SNAPSHOT - have been introduced for upsert enabled tables to ensure consistent query results even when queries and upserts occur concurrently and at very high throughput.

By default, the consistency mode is NONE, meaning the system operates as before. The SYNC mode ensures consistency by blocking upserts while queries execute, guaranteeing that queries always see a stable upserted data view. However, this can introduce write latency. Alternatively, the SNAPSHOT mode creates a consistent snapshot of validDocIds bitmaps for queries to use. This allows upserts to continue without blocking queries, making it more suitable for workloads with both high query and write rates. These new consistency modes provide flexibility, allowing applications to balance consistency guarantees against performance trade-offs based on their specific requirements.

{
  "upsertConfig": {
    "consistencyMode": "SYNC", // or "SNAPSHOT", "NONE"
...
  }
}

For SNAPSHOT mode, one can configure how often the upsert view should be refreshed via a upsertConfig called upsertViewRefreshIntervalMs, which is 3000ms by default. Both the write and query threads can refresh the upsert view when it gets stale according to this config. Changing this config requires server restarts.

One can further adjust the view's freshness during query time without restarting servers via a query option called upsertViewFreshnessMs . By default, this query option matches with that upsertConfig upsertViewRefreshIntervalMs , but if a query sets it to a smaller value, the upsert view may get refreshed sooner for the query; and if set to 0, the query simply forces to refresh upsert view every time.

For debugging purposes, there's a query option called skipUpsertView. If set to true, it bypasses the consistent upsert view maintained by SYNC or SNAPSHOT modes. This effectively executes the query as if it were in NONE mode.

Use strictReplicaGroup for routing

The upsert Pinot table can use only the low-level consumer for the input streams. As a result, it uses the partitioned replica-group assignment implicitly for the segments. Moreover, upsert poses the additional requirement that all segments of the same partition must be served from the same server to ensure the data consistency across the segments. Accordingly, it requires to use strictReplicaGroup as the routing strategy. To use that, configure instanceSelectorType in Routing as the following:

{
  "routing": {
    "instanceSelectorType": "strictReplicaGroup"
  }
}

Enable validDocIds snapshots for upsert metadata recovery

Upsert snapshot support is also added in release-0.12.0. To enable the snapshot, set the enableSnapshot to true. For example:

{
  "upsertConfig": {
    "mode": "FULL",
    "enableSnapshot": true
  }
}

Upsert maintains metadata in memory containing which docIds are valid in a particular segment (ValidDocIndexes). This metadata gets lost during server restarts and needs to be recreated again. ValidDocIndexes can not be recovered easily after out-of-TTL primary keys get removed. Enabling snapshots addresses this problem by adding functions to store and recover validDocIds snapshot for Immutable Segments

The snapshots are taken on every segment commit to ensure that they are consistent with the persisted data in case of abrupt shutdown. We recommend that you enable this feature so as to speed up server boot times during restarts.

Enable preload for faster server restarts

Upsert preload feature can make it faster to restore the upsert states when server restarts. To enable the preload feature, set the enablePreload to true. To enable preloading, enableSnapshot: true should also be set in the table config. For example:

{
  "upsertConfig": {
    "mode": "FULL",
    "enableSnapshot": true,
    "enablePreload": true
  }
}

Under the hood, it uses the validDocIds snapshots to identify the valid docs and restore their upsert metadata quickly instead of performing a whole upsert comparison flow. The flow is triggered before the server is marked as ready, after which the server starts to load the remaining segments without snapshots (hence the name preload).

The feature also requires you to specify pinot.server.instance.max.segment.preload.threads: N in the server config where N should be replaced with the number of threads that should be used for preload. It's 0 by default to disable the preloading feature.

Handle out-of-order events

There are 2 configs added related to handling out-of-order events.

dropOutOfOrderRecord

To enable dropping of out-of-order record, set the dropOutOfOrderRecord to true. For example:

{
  "upsertConfig": {
    ...,
    "dropOutOfOrderRecord": true
  }
}

This feature doesn't persist any out-of-order event to the consuming segment. If not specified, the default value is false.

• When false, the out-of-order record gets persisted to the consuming segment, but the MetadataManager mapping is not updated thus this record is not referenced in query or in any future updates. You can still see the records when using skipUpsert query option.

• When true, the out-of-order record doesn't get persisted at all and the MetadataManager mapping is not updated so this record is not referenced in query or in any future updates. You cannot see the records when using skipUpsert query option.

outOfOrderRecordColumn

This is to identify out-of-order events programmatically. To enable this config, add a boolean field in your table schema, say isOutOfOrder and enable via this config. For example:

{
  "upsertConfig": {
    ...,
    "outOfOrderRecordColumn": "isOutOfOrder"
  }
}

This feature persists a true / false value to the isOutOfOrder field based on the orderness of the event. You can filter out out-of-order events while using skipUpsert to avoid any confusion. For example:

select key, val from tbl1 where isOutOfOrder = false option(skipUpsert=false)

Use custom metadata manager

Pinot supports custom PartitionUpsertMetadataManager that handle records and segments updates.

{
  "upsertConfig": {
    "metadataManagerClass": org.apache.pinot.segment.local.upsert.CustomPartitionUpsertMetadataManager
  }
}

Adding custom upsert managers

You can add custom PartitionUpsertMetadataManager as follows:

• Create a new java project. Make sure you keep the package name as org.apache.pinot.segment.local.upsert.xxx

• In your java project include the dependency

<dependency>
  <groupId>org.apache.pinot</groupId>
  <artifactId>pinot-segment-local</artifactId>
  <version>1.0.0</version>
 </dependency>

include 'org.apache.pinot:pinot-common:1.0.0'

• Add your custom partition manager that implements PartitionUpsertMetadataManager interface

//Example custom partition manager

class CustomPartitionUpsertMetadataManager implements PartitionUpsertMetadataManager {}

• Add your custom TableUpsertMetadataManager that implements BaseTableUpsertMetadataManager interface

//Example custom table upsert metadata manager

public class CustomTableUpsertMetadataManager extends BaseTableUpsertMetadataManager {}

• Place the compiled JAR in the /plugins directory in pinot. You will need to restart all Pinot instances if they are already running.

• Now, you can use the custom upsert manager in table configs as follows:

{
  "upsertConfig": {
    "metadataManagerClass": org.apache.pinot.segment.local.upsert.CustomPartitionUpsertMetadataManager
  }
}

⚠️ The upsert manager class name is case-insensitive as well.

Upsert table limitations

There are some limitations for the upsert Pinot tables.

• The upsert feature is supported for Real-time tables only, and not for Hybrid or Offline tables.

• The star-tree index cannot be used for indexing, as the star-tree index performs pre-aggregation during the ingestion.

• Unlike append-only tables, out-of-order events (with comparison value in incoming record less than the latest available value) won't be consumed and indexed by Pinot partial upsert table, these late events will be skipped.

Best practices

Unlike other real-time tables, Upsert table takes up more memory resources as it needs to bookkeep the record locations in memory. As a result, it's important to plan the capacity beforehand, and monitor the resource usage. Here are some recommended practices of using Upsert table.

Create the topic/stream with more partitions.

The number of partitions in input streams determines the partition numbers of the Pinot table. The more partitions you have in input topic/stream, more Pinot servers you can distribute the Pinot table to and therefore more you can scale the table horizontally. Do note that you can't increase the partitions in future for upsert enabled tables so you need to start with good enough partitions (atleast 2-3X the number of pinot servers)

Memory usage

Upsert table maintains an in-memory map from the primary key to the record location. So it's recommended to use a simple primary key type and avoid composite primary keys to save the memory cost. Beware when using JSON column as primary key, same key-values in different order would be considered as different primary keys. In addition, consider the hashFunction config in the Upsert config, which can be UUID, MD5 or MURMUR3.

If your primary key column is a valid UUID and you are running out of memory due to a high number of primary keys, the UUID hash function can lower memory requirements by up to 35% without bringing in any hash collision risks. If the primary key is not a valid UUID, this hash function stores the primary key as is and skips the UUID based compression.

MD5 and MURMUR3 can also help lower memory requirements. They work for all types of primary key values, but bring in a small risk of hash collision. The generated hash from MD5 and MURMUR3 is a 128-bit hash, so this is beneficial when your primary key values are larger than 128-bits.

Monitoring

Set up a dashboard over the metric pinot.server.upsertPrimaryKeysCount.tableName to watch the number of primary keys in a table partition. It's useful for tracking its growth which is proportional to the memory usage growth. **** The total memory usage by upsert is roughly (primaryKeysCount * (sizeOfKeyInBytes + 24))

Capacity planning

It's useful to plan the capacity beforehand to ensure you will not run into resource constraints later. A simple way is to measure the rate of the primary keys in the input stream per partition and extrapolate the data to a specific time period (based on table retention) to approximate the memory usage. A heap dump is also useful to check the memory usage so far on an upsert table instance.

Example

Putting these together, you can find the table configurations of the quick start examples as the following:

{
  "tableName": "upsertMeetupRsvp",
  "tableType": "REALTIME",
  "tenants": {},
  "segmentsConfig": {
    "timeColumnName": "mtime",
    "retentionTimeUnit": "DAYS",
    "retentionTimeValue": "1",
    "replication": "1"
  },
  "tableIndexConfig": {
    "segmentPartitionConfig": {
      "columnPartitionMap": {
        "event_id": {
          "functionName": "Hashcode",
          "numPartitions": 2
        }
      }
    }
  },
  "instanceAssignmentConfigMap": {
    "CONSUMING": {
      "tagPoolConfig": {
        "tag": "DefaultTenant_REALTIME"
      },
      "replicaGroupPartitionConfig": {
        "replicaGroupBased": true,
        "numReplicaGroups": 1,
        "partitionColumn": "event_id",
        "numPartitions": 2,
        "numInstancesPerPartition": 1
      }
    }
  },
  "routing": {
    "segmentPrunerTypes": [
      "partition"
    ],
    "instanceSelectorType": "strictReplicaGroup"
  },
  "ingestionConfig": {
    "streamIngestionConfig": {
      "streamConfigMaps": [
        {
          "streamType": "kafka",
          "stream.kafka.topic.name": "upsertMeetupRSVPEvents",
          "stream.kafka.decoder.class.name": "org.apache.pinot.plugin.inputformat.json.JSONMessageDecoder",
          "stream.kafka.consumer.factory.class.name": "org.apache.pinot.plugin.stream.kafka20.KafkaConsumerFactory",
          "stream.kafka.zk.broker.url": "localhost:2191/kafka",
          "stream.kafka.broker.list": "localhost:19092"
        }
      ]
    }
  },
  "upsertConfig": {
    "mode": "FULL",
    "enableSnapshot": true,
    "enablePreload": true
  },
  "fieldConfigList": [
    {
      "name": "location",
      "encodingType": "RAW",
      "indexType": "H3",
      "properties": {
        "resolutions": "5"
      }
    }
  ],
  "metadata": {
    "customConfigs": {}
  }
}

{
  "tableName": "upsertPartialMeetupRsvp",
  "tableType": "REALTIME",
  "tenants": {},
  "segmentsConfig": {
    "timeColumnName": "mtime",
    "retentionTimeUnit": "DAYS",
    "retentionTimeValue": "1",
    "replication": "1"
  },
  "tableIndexConfig": {
    "segmentPartitionConfig": {
      "columnPartitionMap": {
        "event_id": {
          "functionName": "Hashcode",
          "numPartitions": 2
        }
      }
    },
    "nullHandlingEnabled": true
  },
  "instanceAssignmentConfigMap": {
    "CONSUMING": {
      "tagPoolConfig": {
        "tag": "DefaultTenant_REALTIME"
      },
      "replicaGroupPartitionConfig": {
        "replicaGroupBased": true,
        "numReplicaGroups": 1,
        "partitionColumn": "event_id",
        "numPartitions": 2,
        "numInstancesPerPartition": 1
      }
    }
  },
  "routing": {
    "segmentPrunerTypes": [
      "partition"
    ],
    "instanceSelectorType": "strictReplicaGroup"
  },
  "ingestionConfig": {
    "streamIngestionConfig": {
      "streamConfigMaps": [
        {
          "streamType": "kafka",
          "stream.kafka.topic.name": "upsertPartialMeetupRSVPEvents",
          "stream.kafka.decoder.class.name": "org.apache.pinot.plugin.inputformat.json.JSONMessageDecoder",
          "stream.kafka.consumer.factory.class.name": "org.apache.pinot.plugin.stream.kafka20.KafkaConsumerFactory",
          "stream.kafka.zk.broker.url": "localhost:2191/kafka",
          "stream.kafka.broker.list": "localhost:19092"
        }
      ]
    }
  },
  "upsertConfig": {
    "mode": "PARTIAL",
    "partialUpsertStrategies": {
      "rsvp_count": "INCREMENT",
      "group_name": "UNION",
      "venue_name": "APPEND"
    }
  },
  "fieldConfigList": [
    {
      "name": "location",
      "encodingType": "RAW",
      "indexType": "H3",
      "properties": {
        "resolutions": "5"
      }
    }
  ],
  "metadata": {
    "customConfigs": {}
  }
}

Quick Start

To illustrate how the full upsert works, the Pinot binary comes with a quick start example. Use the following command to creates a real-time upsert table meetupRSVP.

# stop previous quick start cluster, if any
bin/quick-start-upsert-streaming.sh

You can also run partial upsert demo with the following command

# stop previous quick start cluster, if any
bin/quick-start-partial-upsert-streaming.sh

As soon as data flows into the stream, the Pinot table will consume it and it will be ready for querying. Head over to the Query Console to check out the real-time data.

For partial upsert you can see only the value from configured column changed based on specified partial upsert strategy.

An example for partial upsert is shown below, each of the event_id kept being unique during ingestion, meanwhile the value of rsvp_count incremented.

To see the difference from the non-upsert table, you can use a query option skipUpsert to skip the upsert effect in the query result.

FAQ

Can I change primary key columns in existing upsert table?

Yes, you can add or delete columns to primary keys as long as input stream is partitioned on one of the primary key columns. However, you need to restart all Pinot servers so that it can rebuild the primary key to record location map with the new columns.