Dune.com Announces Full Migration to DuneSQL, Plans to Retire SparkSQL and Postgres Back‑End
The analytics platform for blockchain data is consolidating its query stack around a new engine, promising faster performance but requiring users to move existing workloads.
Background
A few weeks ago Dune.com elevated DuneSQL to the status of its default query engine. Early adoption figures released by the company show that the new engine now powers roughly 85 % of freshly created queries and accounts for 30 % of all query executions on the platform. The move follows a technical blog post from Dune’s CTO that outlined the strategic rationale behind the shift.
What’s Changing
Dune’s leadership has confirmed a phased decommissioning of the two legacy back‑ends that have powered the service until now:
| Legacy System | Current Role | Sunset Timeline |
|---|---|---|
| SparkSQL | Primary query engine for most users until recently | Fully retired by 15 July 2023 |
| Ethereum V1 (Postgres) | Original data store for on‑chain data | Retired alongside SparkSQL |
Both systems were deemed sub‑optimal for handling the unique characteristics of cryptocurrency datasets—high‑volume, rapidly evolving, and often requiring complex joins. Consolidating on DuneSQL is intended to streamline the user experience, focus engineering resources, and improve overall stability and performance.
Migration Support
Recognizing that the transition may be disruptive, Dune is rolling out a suite of assistance tools and enhancements:
- Automated migration assistant – A tool, powered by large‑language‑model (LLM) technology, that can translate existing SparkSQL or Postgres queries into DuneSQL syntax.
- Incremental performance upgrades – Ongoing work to reduce query timeouts and improve execution speed.
- Scheduling and materialized views – New capabilities that let users build reliable data pipelines, mitigating the risk of long‑running queries.
- Spellbook integration – The popular analytics UI, currently running on SparkSQL, will receive a DuneSQL‑compatible version. The existing Spellbook will stay operational until the new release is ready, after which Dune will handle the migration automatically.
Impact on the Ecosystem
The consolidation has several implications for developers, analysts, and the broader DeFi community:
- Unified query language – A single, optimized engine reduces the learning curve and eliminates the need to maintain multiple code bases.
- Potential short‑term friction – Users with extensive legacy queries may need to adjust syntax or restructure queries, especially those that relied on Postgres‑specific functions.
- Performance gains – Early benchmarks suggest DuneSQL delivers lower latency and higher throughput, which could accelerate research and real‑time dashboards.
- Reduced maintenance overhead – By retiring SparkSQL and the old Postgres cluster, Dune can allocate more resources to feature development and infrastructure scaling.
Community Feedback
Dune’s communication acknowledges the inconvenience of the migration and offers a direct support channel via email. The company has apologized for any disruption and framed the effort as the “final migration,” positioning DuneSQL as the definitive solution for crypto data analytics.
Key Takeaways
- DuneSQL is now the default engine on Dune.com, already handling the majority of new queries.
- Legacy engines (SparkSQL, Ethereum V1 Postgres) will be fully retired by mid‑July 2023.
- Migration tools—including an LLM‑driven query converter—are being provided to ease the transition.
- Spellbook, a widely used analytics interface, will receive a DuneSQL‑compatible update; the current version remains functional until the new release.
- Performance and stability are the primary motivations, aiming to make Dune the premier platform for blockchain data analysis.
Developers and analysts are encouraged to begin testing their existing queries against DuneSQL today, leverage the migration assistant where possible, and reach out to Dune’s support team with any concerns. As the ecosystem moves toward a single, high‑performance query engine, the expectation is that data‑driven insights in DeFi and broader crypto markets will become faster and more reliable.
Source: https://dune.com/blog/dunesql-migration
