Dremio Reviews

As a BI developer, my main use case for Dremio involves registering data zones and also doing inquiries, views, and making it persistent to use on BI tools, like Tableau reports. In a project, we create data zones, typically one that we have in our company to receive files. After doing the ingestion inside Dremio, what we do is set up permissions, and after setting up the permissions, the refresh session allows us to go over and start building the views and all those things to be possible to connect via JDBC driver and consume on Dremio dashboard.

The first feature that stands out for me in Dremio is the federated type of query, which allows the possibility to use multiple endpoints without worrying about writing custom SQL that runs only for SQL Server or for Postgres and Redshift. We build a SQL on Dremio standard, and then Dremio goes over and rewrites the whole lineage to make it possible to use multiple sources of data.

The feature impacts our daily work by reducing complexity because we don't need to worry about where the data comes from. We are slightly migrating to Snowflake, so Dremio is currently being replaced because Snowflake is a more robust platform, but we are kind of happy with the work Dremio does.

I think the reduction of complexity is a positive impact that Dremio has had on my organization. The main thing is that it is a place where you can write down simple SQLs and see the lineage, the way it integrates together; this is the best part of it.

We also have a close relationship with the team that does the Dremio maintenance for the database, like upgrading the versions and they know about some specific problems we had in the past, such as a memory leak. We had a memory leak on some versions, which sometimes stopped the service. Since we are using Dremio installed like a server, not a SaaS solution, many times we need to stop and restart the service to clear all the cache and all that, and this is the thing I should add.

I see that many times the new versions of Dremio have not fixed old bugs, and in some new versions, old problems that were previously fixed come back again, so I think the upgrade part could use improvement.

I remember using some features in the past, like pivot tables, which proved to be really difficult, but I know this is a fault also for other vendors. Pivoting, transposing, and unpivoting are often not so good. CTEs also many times prove to be not so good, so I think these two main items could be improved significantly if they standardize them.

I have been using Dremio for approximately three years, since January of 2022.

We had a memory leak on some versions, which sometimes stopped the service. Many times, we need to stop and restart the service to clear all the cache and all that. I rate Dremio a nine in terms of stability. I think it is stable, but we need to restart it many times, and we need to monitor it regularly.

Dremio's scalability can handle growing data and user demands easily.

We have had to reach out for customer support many times, and they respond, so they are pretty supportive about some long-term issues. Many times it's with vendors where there is a missing feature or ongoing problem that they say will be fixed in the next release, but this happens not only with Dremio but with almost every vendor we have.

Neutral

We haven't previously used a different solution before Dremio. Snowflake is the new solution we are currently using to replace Dremio.

I cannot share the exact metric itself, but Dremio surely saves time, reduces costs, and all those things because we don't have to worry so much about the infrastructure to make the different tools communicate. Dremio was a solution that made it possible for us to have different data sources with only one registration on our data lake.

I don't have information about pricing, setup cost, and licensing for Dremio, so I am not entitled to discuss it.

I probably don't remember all the options we evaluated, but we talked a lot about Databricks and Redshift, and Dremio was the best choice. Also, we don't use Tableau Prep because we use Dremio, which is an important thing.

My advice to others looking into using Dremio is that it is a great tool because it keeps all your efforts together. It's a good thing to have if you want to have a unified catalog or metadata or something like this, so this is certainly a good tool to consider. I would rate this review an eight out of ten overall.

I have been using Dremio for a year and a half. My main use case for Dremio is that I am able to access multiple databases and I can easily and quickly connect Dremio with my dashboards.

In my recent project, my databases are full of information about our insurance operation, and I monitor my KPIs by connecting Dremio to my dashboards. I have multiple KPIs that I have to stay tuned to, and I need to access that very quickly and in an easy way, which is how Dremio helps me.

About my use case with Dremio, I find databases that I don't know very much about, and in that web environment of Dremio, I am able to scroll and navigate through the databases by searching for an example or a column name, so it is much easier to navigate through with that kind of help.

The best features Dremio offers, in my opinion, probably include the web environment.

What stands out to me about the web environment is both the interface and the speed; it helps a lot.

In terms of features, querying is very easy and the integration with multiple dashboards is amazing.

Dremio has positively impacted my organization as nowadays we are connected to multiple databases from multiple environments, multiple APIs, and applications, and Dremio organizes everything in an amazing way for me.

A specific outcome that shows how Dremio has helped my organization is definitely better data access.

I wouldn't say there is anything Dremio can be improved on. If I could change something, I would say many developers and programmers, when they are starting to work in this specific field or area, are much more used to SQL Server, the Microsoft way of querying, and Dremio has some features that are different when we are talking about the syntax of coding, so I would improve that.

I have been working with data analysis for five years now.

In my experience, Dremio is stable.

I have never had any experience scaling up or down with Dremio.

I have never tried to reach customer support.

Negative

I didn't get to choose the solution I used before Dremio, which was SQL Server from Microsoft. Nowadays I see that SQL Server solution is much more simple, and Dremio helps me a lot to move beyond that simplicity.

My advice for others looking into using Dremio is to explore the web environment and stay sharp with the ODBC connectors and drivers, as that is very important. I would rate this product a 9.

I use Dremio for proof of concept purposes. I haven't used it in a real-time project, however, I explore Dremio as a data virtualization application in the ecosystem. It is relatively new, possibly a one-year or two-year-old system.

It's almost similar, yet it's better than Starburst in spinning up or connecting to the new source since it's on SaaS. It is a similar experience between the based application and cloud-based application. You just get the source, connect the data, get visualization, get connected, and do whatever you want. 

They say data reflection is one way where they do the caching and all that. Starburst also does the caching. In Starburst, you have a data product. Here, the data product comes from a reflection perspective. The y are working on a columnar memory map, columnar computation. That will have some edge in computation.

They need to have multiple connectors. Starburst is rich in connectors, however, they are lacking Salesforce connectivity as of today. They don't have Salesforce connectivity. However, Starburst does. Starburst has all these capabilities. Dremio has only 15 to 20 connectors, however, Starburst comes with around 50 now.

I have used it for just one month for proof of concept purposes.

I cannot comment on stability as I just worked with it for one month. I haven't worked with large data. When I worked with small data, it was fine at that time.

Internally, if it's on Docker or Kubernetes, scalability will be built into the system. In the SaaS, I'm unsure as I haven't set it up. I don't know how the integrated SaaS works inside. If it were an enterprise setup like Starburst, I know how it works since I have worked there, using Kubernetes, Docker, and everything. I'm not familiar with Dremio's backend, however, it also works on Kubernetes and similar technologies. Hopefully, scalability will be there for sure.

It was just proof of concept, and we were just exploring the product. We did not deal with technical support 

Neutral

It is a SaaS, so it is straightforward to set up.

Regarding features, I'm not sure if they have all the tools like data governance, data quality, and data lineage integrated. If not, they need to build those tools as well to check the data quality and lineage. Data discovery is there. Connectivity-wise, Starburst is way better, however, Dremio might have a better computing path, possibly delivering data faster than Starburst. No direct comparison can be made, so I cannot comment further. 

Overall, you can rate it as eight out of ten.

I have been using Dremio on and off as a data warehouse for the past three years.

My main use case for Dremio is that we use it as a logical data warehouse where we use Dremio with VDSs as an alternative to AWS Glue or Apache Hive. As we are working with our ETL at the end of all of it, after the data types and everything have been cast, we make that available on Dremio as VDS and then we move on to our further data warehousing schemes within Dremio.

We use Dremio enterprise-wide now, and the key use case has been reducing our costs when it comes to data storage.

Our main use case for Dremio is as a data warehouse, and the challenge that it helped us solve is that physical data warehouses such as Redshift have storage and hardware upscaling conflicts. Dremio helps us decouple those and lets us catalog more. We can manage everything under one system.

The best features Dremio offers include having a single system where we can manage all of our data cataloging and visualization or virtualization.

The interface is a plus over the traditional warehousing solutions, which makes it easier to work with Dremio compared to other solutions I've used.

Having everything under one system and an easier-to-work-with interface, along with having API integrations, adds significant value to working with Dremio.

Dremio has positively impacted my organization by helping us create a single source of truth, a singular data warehouse where we can have access to all of the data sets. The fact that Dremio has a clear role-based access management system helps us significantly, as we can have roles segregating all of the data, while users with the appropriate roles can access everything.

Dremio could be improved by making it easier for data cataloging, especially when working with open table formats, as you have to choose a data format and then go into it. With the Dremio software version that we're using, all that requires a learning curve, and only when you go to the premium cloud version do you get Dremio Arctic. It should be easier to get Arctic or an open-source version of Arctic onto the software version so that development teams can experiment with it.

I have been working in my current field for five years.

In my experience, Dremio is reasonably stable.

I haven't pushed Dremio's scalability to its limit, so I cannot provide detailed information about that.

I haven't had the need to interact with Dremio's support team yet.

Positive

Before Dremio, we used Hive and AWS Glue, and we wanted to shift to an open-source version that is newer than Hive and provides flexibility in moving between cloud and on-premises.

We evaluated other options before choosing Dremio, including Presto and Trino, but we did not find reasonable advantages to using them.

I don't have the appropriate information regarding pricing, setup cost, and licensing as that is managed by a different team.

I cannot share return on investment information from using Dremio.

We evaluated other options before choosing Dremio, including Presto and Trino, but we did not find reasonable advantages to using them.

I would advise others looking into using Dremio to study the tool beforehand. Dremio has several different offerings, and the best way to get into it is to use the open-source on-premises version to experiment with it. To extract the complete power of the platform, organizations should educate themselves with the complete information and compare it with other solutions since a single solution cannot fit everywhere. Educating the team before adopting a technology is better than just adopting a suggested package. I rate Dremio 8 out of 10.

I have been using Dremio to select multiple data sources, retrieve multiple data sources, and create and maintain Iceberg tables for the data lakehouse.

My main use case for Dremio is to actively create queries for multiple data sources. For example, I work with data marts and databases. I also create new tables on the data lakehouse and perform cross-joins from multiple data sources.

One project was to predict income, and I used multiple income tables from the data warehouse. I created pipelines from the data warehouse to the data lakehouse and selected all the old tables using Dremio.

Dremio has positively impacted my organization by creating faster development and POC projects. I create the data lakehouse and develop lakehouse tables, for example, creating Iceberg tables.

I have noticed time saved, improved collaboration, and improved multiple data connections.

The best features Dremio offers in my experience include native data catalogs because my company didn't create new data catalogs, but Dremio can use a native data catalog.

The native data catalog feature has made the development cycles faster. Multiple data sources can be queried in one query, and it is not just a query engine.

I think Dremio can be improved to query faster than multiple sources and big data lakehouses. The community version could be updated for new data catalog connections.

I think the documentation is sufficient. For support, the Enterprise version is adequate.

I have been actively using Dremio for one year in my work, and I have used Dremio Enterprise for six months.

Dremio is stable in my experience.

Dremio's scalability is good, but I use on-premises models, so I do not have experience with the cloud.

Customer support is great. I would rate the customer support an eight.

I worked on Trino before Dremio, but it is a query engine. I needed other techniques for how it works.

I chose eight out of ten because Dremio is expensive for the Enterprise version, particularly regarding licensing.

Trino was the only option I evaluated because I need to connect multiple tools.

Regarding Dremio's AI capabilities, I think its governance and security is good, but I did not use Dremio's AI because my company's security is high.

I do not have enough experience with Dremio's AI capabilities to comment on its accuracy and reliability of output.

I would advise using Dremio if you have a small project using a data lakehouse or if you need to connect multiple data sources. You might consider switching from Trino to Dremio. I would rate this product an eight overall.

We use Dremio for financial data analytics and as a data lake. We connect Dremio with Oracle, Docker, MySQL, and utilize it for Power BI. 

Additionally, we use it to process data from MongoDB, although we face occasional challenges with NoSQL integration.

Dremio is very easy to use for building queries. It's easy to connect Dremio to various databases and data sources like Oracle and MySQL. It is also very flexible, providing us with scalability and integration capabilities effortlessly.

There are performance issues at times due to our limited experience with Dremio, and the fact that we are running it on single nodes using a community version. We face certain issues when connecting Dremio to MongoDB, especially with max values, which seem to be inconsistent in Dremio. 

Additionally, licensing is quite expensive, and we feel the need for more flexible schema capabilities, especially in embedding JSON from MongoDB.

We have been using Dremio for more than two years.

In terms of stability, we experience performance issues occasionally, partly due to our limited experience.

We have not yet scaled Dremio because we are using the community version, and require a license for scaling. We need to learn how to scale up and out more effectively.

We haven't submitted any questions for technical support yet.

Positive

We have worked with various data warehouse solutions, like Oracle, and also used databases on Azure.

The initial setup was straightforward. We run Dremio on an Ubuntu dedicated server on-premises, and it took us about a day to set up.

We see savings because we don't need more personnel to develop and maintain Dremio. It's cost-effective in terms of manpower.

The licensing is very expensive. We need a license to scale as we are currently using the community version.

We used Oracle and databases on Azure before using Dremio.

Dremio is very flexible and easy to use, making it very suitable for our team. I would recommend Dremio for similar use cases due to its flexibility. 

On a scale of one to ten, I would rate Dremio at eight.

We have been using it to build one of our frameworks. We primarily use Dremio to create a data framework and a data queue. It's being used in combination with DBT and Databricks.

We're still in the exploration phase with Dremio, so it's a bit early to determine its most valuable feature. We're currently deploying it across different departments for various use cases and learning from these internal applications.

We've faced a challenge with integrating Dremio and Databricks, specifically regarding authentication. It is not shaking hands very easily. We had to set up two different VMs and execute them in a different manner and integrate them. 

I've been using Dremio for about two to three months now. However, one of our teams has been using it for the past year.

From my three months of experience, I haven't noticed any stability issues with Dremio.

In my department, which focuses on data and AI, we have about 538 people. I'm not sure how many are actively using Dremio.

The installation process was quite smooth and didn't present any issues.

We currently have Dremio on the cloud. For proof of concept (POC) purposes, we are using it on-premises.

We are currently evaluating Dremio against other similar products. But at first glance, I would recommend using Dremio.

Considering my limited access and experience over these three months, I would rate Dremio around a seven out of ten.

I can visualize traffic from BI and Tableau on the same page and have my tables and schema on the same page. The data link comprises everything. If I want one structure, I connect it to a big table in the hive and the data team that could read my SQL work on my tables, schemas, table structures and everything all in one place. Dermio is as good as any other Presto engine.

Everyone uses Dremio in my company; some use it only for the analytics function.

The most valuable feature is that you can generate refresh reflections and create your visuals (VDS) because it makes it easier to monitor day-to-day data structure. I can use Dermio to create a visual table without impacting the original by creating an opportunity on my own. I can work on my videos and create reflections, and Dremio allows me to use those reflections. I can know the health of my tables, whether healthy or unhealthy, daily. It is one of the most valuable features of Dremio.

One of the areas of improvement is that a table does not break and shows errors. When the 23rd version was released, I had to contact Dermio customer support, and they suggested I update the database to run the table. Sometimes you face common errors like tables running strictly and taking time to run. I think Dermio can improve this part.

An additional feature can be a feature where everyone can see the tables.

I have been using Dremio for over 3years now.

I would rate the scalability an 9/10 because it shows errors sometimes.

I would rate the scalability a  10/10 based on its recent enhancements.

We had Dremio support service at every point of time we requested it.

Positive

Previously, I used Presto regularly.

The initial setup was straightforward.

The deployment was in-house, and only two to three people were involved. We had direct Dermio support communication to support us. We have both cloud and in-house deployment methods.

With Dremio its not a loss

Every tool has a value based on its intended purpose and use, and the pricing is worth its value.

We tested some other options but Dremio proofs to be better in terms of reliability and scalability, so we submitted our reports and reviews.

I suggest that you give it a try and see for yourself. Dremio is as fast as Presto. For example, ten billion rows can return in less than five seconds.

I would rate it a 10 out of 10.

We use Dremio for data engineering.

Dremio has resolved my data lineage and data governance problems. The solution has also resolved the data availability for a different range of users, which used to be a problem.

Dremio allows querying the files I have on my block storage or object storage. The solution gives me a place where I can play around with the data virtually by creating VDSs or PDSs. Dremio works just like a data warehouse on top of my data lake, which is interesting.

Dremio's interface is good, but it has a few limitations. I cannot do a lot of things with ANSI SQL or basic SQL. I cannot use the recursive common table expression (CTE) in Dremio because the support page says it's currently unsupported.

The use case I am working on requires building trees and hierarchical structures. Most of the time, it requires complex nested data structures to be made simpler for end users. It would be good if Dremio could provide a way to create trees just like Oracle does using commands like CONNECT BY and NO CYCLE.

You can use a few languages to simplify complicated JSON and XML. It would be very helpful if Dremio could provide a solution to simplify building trees and building meaningful data from complex data.

I rate Dremio ten out of ten for stability.

I rate Dremio ten out of ten for scalability.

I rate Dremio ten out of ten for the ease of its initial setup.

We implemented the solution through an in-house team. Dremio's deployment can be done quickly.

Overall, I rate Dremio ten out of ten.

Dremio is a platform that enables you to perform high-performance queries from a data lake. It helps you manage that data in a sophisticated way. The use cases are broad, but it allows you to make extremely good use of data in a data lake. It gives you data warehouse capabilities with data lake data.

Dremio enables you to manage changes more effectively than any other data warehouse platform. There are two things that come into play. One is data lineage. If you are looking at data in Dremio, you may want to know the source and what happened to it along the way or how it may have been transformed in the data pipeline to get to the point where you're consuming it. 

There's another thing called data providence. They're tied together. Data providence allows you to go back and recreate the data at any particular point in time. It's extremely important for compliance and governance issues because data changes all time. How did it change? What was it three days or months ago? You may have made some decisions based on data that was three months old, so you might need to revisit those.

It's essential for things like machine learning and deep learning, where you are generating AI models off data. When the model stops working or doesn't work as expected, you need to figure out why. You have to go back and adjust the datasets used to train the model. We do that through an open-source project called Nessie, which is their basis for providing data lineage and data province capabilities. It's super powerful.

Arrow is another open-source project for storing data in memory and performing data query operations. Data sits on a disk in one format. If you want to do anything with data, you have to load it into your computer and put it into memory so you can work with it. Arrow provides a format in memory that enables the whole library to perform various operations on that data. Every vendor has its own way of representing data in memory.

They've latched onto an industry standard and developed it so it's open. Now people can use the exact same format in memory to do operations and use the library set to perform functions on data. New developers can decide if they want to develop their own memory format or use one that's already there.

Data transfer is a massive problem when you're working with large datasets, doing advanced analytics, and trying to train machine learning or deep learning models. What happens often is companies downsample their data sets to do training on models because transferring and managing data on a deep learning or machine learning platform is too much.

They have an automated tool for building SQL queries, so you don't need to know SQL. That interface works, but it could be more efficient in terms of the SQL generated from those things. It's going through some growing pains. There is so much value in tools like these for people with no SQL experience. Over time, Dermio will make these capabilities more accessible to users who aren't database people.

The core functionality is rock solid in my experience. 

Dremio is designed for massive scale-building.

I rate Dremio eight out of 10. Dremio focuses on enterprise customers, so they have a solid support organization. They have professional services and full support. They have a great customer-centered team, but there are some challenges. For example, they don't have enough staff. As more customers buy Dremio, they need more integrators to help them get started. They can't do it all themselves.

Positive

The initial setup is pretty straightforward. I've set it up in two different ways. The first was a three-node configuration. I used virtual machines to set up three different nodes and cluster data. That worked fairly well. I also created containerized versions using Kubernetes. It was much easier to do that. Dremio also has cloud versions where you don't need to do anything. You can go to the cloud and consume it.

I rate Dermio 10 out of 10. Data lake 2.0 or 3.0 is gradually replacing data warehouses, but it's not quite there yet. We can't get rid of data warehouses, but it's heading in that direction. Dremio is a phenomenal place to start with that mindset shift and architecture shift. Despite the shortcomings, Dremio has made better progress in moving the industry forward than anybody else out there.