Skip to content

Analytics DB

About Analytics DB

Nodely DataWarehouse is based on open source ClickHouse DW.


DW interface is exposed by default as:

Custom endpoints/integrations include

ClickHouse SQL

Clickhouse has powerful SQL language extensions and functions. It pays to learn about the join types, join orders, aggregations.

See: ClickHouse SQL documentation


Allo analytics schema consists of several databases.

Each customer has access to dedicated scratch (no backup) database and common databases each corresponding to the named network - mainnet, voitestnet with custom ledger schema. Scratch database is available only on primary cluster. Failover does not have dedicated scratch database by default. Manual setup of table replication or backup schedule is possible with PS.

[ DW schema](


Dictionaries are in-memory maps that can quickly map between a Uint64 ID and a named value.

Dictionaries periodically refresh data from their source. Refreshes are atomic, no duplicated entries are to be expected.

Dictionaries have dict_ as a prefix in their names and are grouped in databases with _mem suffix:

  • mainnet_mem
  • testnet_mem
  • betanet_mem

Each dictionary has corresponding JOIN Table that can be used for classical (yet instant) joins .

All tables with no dict_ prefix in *_mem databases can be used for instant joins.



All native (non dictionary based) tables are append only and can contain duplicated rows. Deduplication is eventual and nondeterministic.

To see dedupilcated view one must use FINAL keyword or user argMax functions to select only relevant rows.

Working with external data

Working with external data

Default settings for end-users

optimize_move_to_prewhere_if_final1Optimize deduplication by applying WHERE filters before dedup phase.
final1Automatically appliesĀĀ modifier to all tables in a query
union_default_modeDISTINCTDefault union behaviour when not specified

Exposing queries as API

Exposing queries as API

SQL Playground

Allo Algorand BI

ODBC DW Access