Question

I have a CSV file with 74 columns and about 60K rows. The contents of this CSV file have to be imported to a MySQL database, every month.

After the data is inserted, the end-user can query the contents of the MySQL database with predefined filters.

Putting everything in a single table would mean faster inserts, but slower reads. Splitting the content in multiple tables (with foreign keys) would mean slower inserts, faster reads and, I think, higher chance of failure.

What do you think is the best option for me, or are there any other possibilities?

Answer 1

If all the data relationships (between the buses, clients, and trips) are 1 to 1 and information is not being duplicated throughout your CSV, you can go with a single table for these reasons:

1. Simplest conversion from CVS to database, each column in the CVS will correspond with one column in the database
2. Anyone that works on the database after you will know exactly what data is where because it will “look like” the CVS
3. Your main concern “Reads being slower” won’t be a big problem because when you query the database for information, you ask for only the data you want and filter out the columns you don’t. (e.g. SELECT departure, arrival, distance FROM bustrips WHERE distance > 1000)\

However, if you look at the data, and there is a massive amount of duplication in the CVS, possibly from more than one client riding on the same trip, or the same bus is used for more than one trip, etc. I would create a new table for each block of unique data. One example that I might already see would be a new table for buses:

Bus_ID;
Numberplate;
Handicap;
Odometer reading;

I hope this helps you make the decision. It is not about “easy read” vs. “easy write” it is about information clarity by reduction of redundancy.