If you have a single block of data, using fputcsv to create the table works well; it assumes the first row is the header and subsequent rows have the same format.

But what if you have multiple chunks of data in different formats that you want to write to CSV:

Business Name: name-of-business.   Date: current date

Notes: single line of notes


Model | Product 
model1 | Product 1
model2 | Product 2
model3 | Product 3

The first row is basically a single row table with 4 columns and no headers The second row is a single row table with 2 columns and no headers Then comes the data - a typical two-column table with headers