CSV files

A table as text

A CSVCSVComma-separated values: a table saved as plain text, one row per line, with commas between the values. file is a table written as plain text: one rowrowOne line of a table: one record, like one item and its price. per line, with commas between the values. The first row is usually the headerheaderThe first row of a table, naming each column: item,price., naming the columns. Open Files under this runner to see prices.csv, a short Merkato price list:

import csv

with open("prices.csv", newline="", encoding="utf-8") as f:
    rows = list(csv.reader(f))
print(rows)

item,price
ቡና,320
በርበሬ,180
ምስር,95

csv.reader turns each line into a list of strings, header included. Open CSV files with newline="", as the csv documentation asks: the module handles line endings itself. Without it, a quoted value with a line break inside reads wrong, and on Windows every row you write gets an extra blank line after it.

Lists make you remember that the price is position 1. csv.DictReader uses the header instead: each row becomes a dict whose keys are the column names. Predict the last line:

import csv

with open("prices.csv", newline="", encoding="utf-8") as f:
    rows = list(csv.DictReader(f))
print(rows[0])
print(type(rows[0]["price"]))

item,price
ቡና,320
በርበሬ,180
ምስር,95

{'item': 'ቡና', 'price': '320'}
<class 'str'>

If you picked int, you trusted the way 320 looks. A CSV file has no types, only text, so every value comes back as a str. Look for the quotes in '320'.

Totaling a column

Forget that, and adding up the prices fails on the first row:

To add up the prices, convert each one with int() and keep a running total (lesson 3.2). Real files have mistakes, so this one has a price written as 60 birr. Catch the ValueError (lesson 7.2) and skip that row instead of stopping:

import csv

total = 0
with open("prices.csv", newline="", encoding="utf-8") as f:
    for row in csv.DictReader(f):
        try:
            total = total + int(row["price"])
        except ValueError:
            print(f"Skipping {row['item']}: {row['price']!r}")
print("Total:", total)

item,price
ቡና,320
ሻይ,60 birr
በርበሬ,180

total is the running total; !r in the f-string shows the bad value with repr (lesson 6.1), quotes included. Now replace the four try/except lines with just total = total + int(row["price"]) and run it: the whole program stops at the third row with ValueError: invalid literal for int() with base 10: '60 birr', and you never get a total.

Writing a table

csv.DictWriter writes dicts as rows. You give it the column names, write the header once, then one row per dict:

import csv

basket = [{"item": "ቡና", "price": 320}, {"item": "ምስር", "price": 95}]

with open("basket.csv", "w", newline="", encoding="utf-8") as f:
    writer = csv.DictWriter(f, fieldnames=["item", "price"])
    writer.writeheader()
    for row in basket:
        writer.writerow(row)

It prints nothing; look in Files for basket.csv. The numbers became text on the way out. csv.writer does the same for lists: writer.writerow(["ቡና", 320]).

Why not split(',')?

A line of CSV looks easy to cut with split(",") from lesson 6.1. But a value can have a comma inside it, and then CSV puts the value in double quotes. Predict:

line = '"ቡና, ሲዳሞ",320'
print(line.split(","))

['"ቡና', ' ሲዳሞ"', '320']

split cut the item in two and kept the quotes. The csv module reads the quotes the way they're meant:

import csv

line = '"ቡና, ሲዳሞ",320'
print(list(csv.reader([line]))[0])

csv.reader reads any list of lines, not only a file. When it writes a value with a comma, it adds the quotes for you. So let the csv module read and write CSV, and never build or cut the commas yourself.