Although I agree with most of the software developers here, what you heard is not virtueless.
There might be several reasons to why some big companies rely on CSV files instead of databases. Frankly, some of the more traditional industries in healthcare, finance, manufacturing and so on do use CSV files more widely than databases.
1. Distribution of data
A flat CSV file is perhaps the easiest way to share a dataset. A database on the on the other hand requires discrete provisioning of access controls, and privilege separations. It’d cost the IT department a lot of resources to provision granular access controls for each employee who has requested data. Removing, and editing their accesses can be even worse.
The easier thing to do is to open it up on a spreadsheet program, hide a few columns, filter out some rows and then share it to the recipient all without the IT department granting a database access role.
2. Versioning control
Bigger firms typically have some centralized databases that they keep maintaining in a rigorous and efficient manner. It’s hard for the server to handle data request from all employees. They would distribute certain portion of the database every now and then. The employees however need to keep each copy each time. This is why you see people date their CSV files religiously.
This is certainly achievable via SQL database. I know analysts who’s in charge all her department’s SQL queries. She keeps a nice hierarchical structure for these queries in relation to the tables associated.
For people outside of IT or analytics department, a flat file might just do.
3. Reporting
Again, sending tons of queries against a company server can cause problems. There could be delays, congestion, and other problems. Visualizing a flat file on a local computer is a lot easier and robust. I guess this is why employees at bigger companies still share CSV files to each other.
These are the main reasons I can think of. I use a cloud-based data management tool called Acho to turn all my CSV files into databases. It actually gives you the best of two worlds actually. On one hand, you can still keep all your CSV files separated. One the other hand, you can build a SQL database for them all with no coding of course.