I'm deciding between SQL or document-oriented db like Mongo (AWS DocumentDB). In SQL I would plan for each company-time period pair to get its own row i.e. company A Q1 2020 is its own row and each column represents an accounting measure, like revenue. SQL seems like it would perform these calculations fast, but NoSql would allow for less structured data (even within GAAP accounting different names are used for similar accounting measures.)
What are the pros and cons of each? Any better solutions than these?