csv.reader vs csv.DictReader

Question

Accepted Answer

csv.reader treats every row as a plain list: the first row is just another list, and column positions are accessed by integer index. csv.DictReader reads the first row as the header and uses those values as keys. Every subsequent row is an OrderedDict (Python 3.7 and earlier) or a regular dict (Python 3.8+) where you access values by column name. For data work, DictReader is almost always preferable. Column names are self-documenting, your code survives column reordering, and you don't need to remember that price is index 3. The tradeoff is a small overhead from dict construction per row. For files with hundreds of millions of rows where you only need one or two columns, reader with explicit index access can be faster. You can also pass fieldnames= to DictReader to override or supply headers when the file has none.