Overall Satisfaction with Matillion
We used Matillion on a small, 3-person team to manage ingestion and transformation for our data warehouse, pulling from numerous datastores and vendors to a common warehouse schema.
- Shallow learning curve - Matillion is powerful, but only as complicated as you need at a given time.
- Superb helpdesk and documentation.
- Jobs are translated directly to SQL which makes them easy to debug.
- Parsing of flat files is inconsistent depending on the source
- The product is constantly evolving (good!) but that means relearning parts of the interface every few months
- Support for a larger set of APIs and database engines would be helpful
Interface is responsive and straightforward, and very easy to grasp. 9/10 because administrative settings are nonintuitive and the scope of variables takes some getting used to.
With the help of the support team, we had a proof of concept up and running in an afternoon. Since Matillion comes as an AWS package, there was no configuration needed to get the instance spun up, which was fantastic. Connecting to data sources that require their own JDBC drivers took as much as an hour, but settings are portable, so this was a one-off process for each new source.
Resource use can balloon when performing even moderately-complex transformations, and race conditions still have a tendency of appearing when more than a few processes run concurrently.
Matillion is great for a team's first foray into ETL. It holds your hand exactly as much as is needed, and scales brilliantly within the scope of a small team. It becomes less ideal as you add more data sources and processing. It interacts well with Python, but any other scripting is harder to integrate.