dbt vs. erwin Data Modeler vs. Fivetran

The prerequisite is that you have a supported database/data warehouse and have already found a way to ingest your raw data. Then dbt is very well suited to manage your transformation logic if the people using it are familiar with SQL. If you want to benefit from bringing engineering practices to data, dbt is a great fit. It can bring CI/CD practices, version control, automated testing, documentation generation, etc. It is not so well suited if the people managing the transformation logic do not like to code (in SQL) but prefer graphical user interfaces.

Incentivized

I have had a chance to use few other data modeling tools from Quest and Oracle, but I am most comfortable using erwin Data Modeler. They understand your data modeling needs and have designed the software to give you a feeling of completeness when you are designing a data model.

Incentivized

Fivetran's business model justifies the use-case where we require data from a single source basically a lot of data but if the requirement is not on the heavier side, Fivetran comes to costly operation when compared to its peers. Otherwise, I'll recommend Fivetran for stability and update and seamless service provider.

Incentivized

• dbt supports version control through GIT, this allows teams to collaborate and track the data transformation logic.
• dbt allows us to build data models which helps to break complex transformation logic into simple and smaller logic.
• dbt is completely based on SQL which allows data analyst and data engineers to build the transformation logic.
• dbt can be easily integrated with snowflake.

Incentivized

• Reverse Engineering: I love the way we can import an SQL file containing schema meta data and generate ER diagram out of it. This is specifically useful if you are implementing erwin Data Modeler for an existing database.
• Forward Engineering: We use this feature very frequently. Where we do database changes in our physical and logical data models and then generate deployment scripts for the changes made.
• Physical vs Logical Models: I like to have my database model split into physical and logical models and at the same time still linked to each other. Any changes you make to logical model or physical model shows up in the other.

Incentivized

• Easily connects to source data using delivered connectors
• Transforms data into standard models and schemas
• Has very good documentation to help quickly setup connectors

Incentivized

• Field-level lineage (currently at table level)
• Documentation inheritance - if a field is documented the downstream field of the same name could inherit the doc info
• Adding python model support (in beta now)

Incentivized

• Cost is the single largest barrier to entry with this tool.
• Would like to see a greater variety of export options available for use outside of the tool, for example as a vector graphic.
• Support has been uneven at times.

Incentivized

• Very difficult to get connectors enhanced if a specific needed object is not supported by them
• Depending on the edition needed and the data volumes, can get quite expensive

Incentivized

dbt is very easy to use. Basically if you can write SQL, you will be able to use dbt to get what you need done. Of course more advanced users with more technical skills can do more things.

Incentivized

I had a lot of experience using erwin Data Modeler for designing data models. I think it's pretty intuitive and easy to use. It has enough features to represent your database requirements in form of a model.

Incentivized

Very easy and intuitive to setup and maintain as there usually are not that many options. Very well documented (e.g. how to setup each connector, how the schema looks like, any specific features of this connector etc.). Also the operation is intuitive, e.g. you have status pages, log pages, configuration pages etc. for each connector.

Incentivized

It runs pretty well and gets our data from point A to point cluster quickly enough. Honestly, it's not something I think about unless it breaks and that's pretty rare.

Incentivized

CA customer support and our account manager have been able to support us with any issues that we have had, from managing our serial keys to issues we logged tickets to resolve. There are aspects of key management that have made it difficult over the years but support usually has worked with us.

Incentivized

I actually don't know what the alternative to dbt is. I'm sure one must exist other than more 'roll your own' options like Apache Airflow, say, bu tin terms of super easy managed/cloud data transforms, dbt really does seem to be THE tool to use. It's $50/month per dev, BUT there's a FREE version for 1 dev seat with no read-only access for anyone else, so you can always start with that and then buy yourself a seat later.

Incentivized

Not listed, but I've only used alternatives built into something like the Squirrel SQL editor. That one is semi-functional but lacking many features and, in some instances, just plain wrong. The only pro there is that it's freely available and works over ODBC. I've tried some of the other free ones like Creately but didn't have much success.

Incentivized

We never seriously considered using anything else. Our data engineers had used Fivetran extensively in previous roles so when it came time to make a decision, there wasn't much of a process. They gladly signed the contract with Fivetran pretty quickly.

Incentivized

• Simplified our BI layer for faster load times
• Increased the quality of data reaching our end users
• Makes complex transformations manageable

Incentivized

• It is difficult to assign a monetary value to it. The initial learning curve was fairly long.
• It allows us to have better support documentation.
• It allows ease of communications between functional units.

Incentivized

• It has been very positive in serving BI team with new source requests
• It has been OK at scaling to match as data volumes as source data size grows

Incentivized