Data Build Tool, aka DBT, is best used for transformations and is the T in ETL and/or ELT. The “normal” usage of DBT is the T in ELT method. However there is no reason why the DBT tool can not be used in the ELT method.

DBT is very innovative in the way it performs transformation work by splitting up the tasks. This splitting up of tasks is also the reason why it is much faster than Java based tools. It uses SQL for Data queries, YAML (Yet Another Markup Language) for Modelling & SQL Injection and Python for the Platform & Data Exchange between database & nodes. Java based tools are slower because the data pipeline steps are usually compiled to create a single code and there is an overhead of JVM reading that code. For its speed and ease of use - DBT is the underlying language used by cloud databases and analytics platforms like Snowflake and ThoughtSpot.

DBT brings software development lifecycle to data management. DBT automates a lot of tasks that are manual in most tools. Tasks like version control, re-usable named macros, etc. are all automated.

It is an open source language and the programmers love it for its ease of use.

Where DBT may fail:

• It is a command line language and debugging it or explaining the steps to a business can be very time consuming.
• Libraries are being built as DBT is till in its infancy.
• It remains to be seen which version/flavor will be last.
• It is an incomplete tool i.e while DBT open source triggers no licenses - other tools are required for data extraction and loading. An extraction and loading tools could be open source as well but a team to run and maintain a second tool will still be required.
• The entire transformation has to be thought through, else it results in a costly process especially if paying by the transaction as is often the case with cloud databases.

It is open source and therefore costs nothing in licenses.

It is blazing fast (assuming a great programmer).

Great for ad-hoc analysis where an (great) analytics engineer can gather the data and run analysis on their own.