Apache Trino, the open-source, distributed SQL query engine, has been quietly gaining traction in New Zealand’s tech ecosystem—particularly among enterprises and data teams that demand speed, scalability, and cross-database compatibility. While it’s not yet the dominant player in local data infrastructure, its adoption is growing, especially among organisations leveraging multi-cloud or hybrid environments. For those who’ve been waiting for a unified query engine that works across Snowflake, BigQuery, PostgreSQL, and more, Trino offers a compelling alternative to traditional ETL pipelines. Its performance benchmarks—particularly in complex joins and aggregations—often outstrip older solutions, making it a strong contender for modern data workflows.
In this review, we’ll examine Trino’s strengths and weaknesses from a New Zealand perspective, focusing on its real-world utility for local businesses and government agencies. We’ll explore how it integrates with existing data stacks, its performance edge in specific scenarios, and where it may fall short for certain workloads. By the end, you’ll have a clear picture of whether Trino is the right tool for your organisation—or if you should consider alternatives like Presto or Spark SQL instead.
Why New Zealand Businesses Are Watching Trino
Trino’s appeal in Aotearoa stems from several key advantages that align with the challenges faced by local data teams. First, it eliminates the need for multiple query tools by providing a single interface for querying data across disparate sources. This is particularly valuable for organisations that use a mix of cloud and on-premises databases—common in industries like finance, healthcare, and agribusiness, where data silos are a persistent issue. For example, a large Māori-owned bank might use Trino to query both customer transaction data in Snowflake and legacy financial records in PostgreSQL without rewriting queries for each platform.
Second, Trino’s performance metrics in real-world scenarios often rival—or exceed—those of traditional ETL tools. In a recent benchmark conducted by a Wellington-based data consultancy, Trino processed a 1TB dataset containing 100 million rows in under 12 minutes, compared to 25 minutes for a similar setup using Apache Spark. While Spark is often the default choice for large-scale processing, Trino’s strengths lie in its ability to handle complex SQL operations efficiently, which is critical for decision-making tools like reporting dashboards and real-time analytics.
- Trino supports over 200 SQL dialects, including PostgreSQL, MySQL, Hive, and Iceberg, ensuring seamless integration with existing data lakes.
- In New Zealand, Trino’s open-source model reduces licensing costs for organisations that might otherwise pay premium fees for proprietary SQL engines.
- Its distributed architecture scales horizontally, making it ideal for growing datasets—critical for agribusinesses like Fonterra, which manage vast amounts of farm data.
- Trino’s low-latency query performance is particularly notable for time-series data, a common need in energy and logistics sectors.
- Local adoption is growing among government agencies, such as the Ministry for Business, Innovation, and Employment (MBIE), which uses Trino to unify data from multiple sources for policy analysis.
The Trino Experience: Performance and Usability
While Trino’s technical capabilities are impressive, its usability can feel less polished than some alternatives. For instance, setting up a Trino cluster requires careful configuration of network connectivity between nodes, which can be a hurdle for teams unfamiliar with distributed systems. In a practical test, a team at a regional council in Auckland spent an extra three hours troubleshooting network latency issues between their Trino nodes and external data sources. However, once operational, Trino’s query interface is intuitive, with a familiar SQL syntax that most data professionals will recognise.
Performance varies significantly depending on the query complexity. Simple aggregations run smoothly, but very large or nested queries may hit performance limits unless optimised with proper indexing. A case study from a Tauranga-based fintech firm found that their Trino queries for customer churn analysis improved from 45 minutes to under 5 minutes after implementing materialised views and query partitioning. This highlights the importance of pre-optimisation—something that might not be as critical for smaller datasets but becomes essential as workloads scale.
Trino vs. the Competition: Where It Shines and Where It Falls Short
When comparing Trino to other SQL query engines, its strengths lie in its cross-database compatibility and performance in complex queries. For example, it outperforms Presto in scenarios involving joins across multiple cloud platforms, making it a better fit for organisations like KiwiRail, which manages data across AWS, Azure, and on-premises systems. However, Trino’s lack of built-in machine learning capabilities means it’s not ideal for teams that rely heavily on predictive analytics—where tools like Apache Spark or Databricks might be more suitable.
Another limitation is Trino’s limited support for certain advanced SQL features compared to newer engines like Apache Iceberg. While this isn’t a dealbreaker for most organisations, it could be a factor for those working with very large-scale datasets that require advanced partitioning strategies. That said, Trino’s growing ecosystem of connectors and plugins means it’s catching up quickly in this area, with new features being added regularly.
The Future of Trino in New Zealand
The adoption of Trino in Aotearoa is still in its early stages, but its potential is undeniable. As more organisations move towards data-driven decision-making and multi-cloud architectures, Trino’s ability to unite disparate data sources will become increasingly valuable. For instance, the New Zealand Government’s push towards open data could see Trino adopted more widely across agencies, as it provides a cost-effective way to query and analyse public datasets.
For businesses looking to future-proof their data infrastructure, Trino offers a compelling alternative to traditional ETL pipelines. Its open-source nature also aligns with New Zealand’s growing focus on innovation and cost efficiency in the tech sector. While it may not replace Spark or Snowflake entirely, it’s a tool that should be on the radar for any organisation serious about optimising their SQL-based data workflows. As the technology continues to mature, we expect to see even more local organisations embracing Trino as a key component of their data stack.
