Clean and validate CSV rows

csv.DictReader returns strings for field values. Convert minutes explicitly and attach the row number to errors so a user can find the source.

For file practice, download the Python practice files v1.0.0 bundle. The examples write their own files, so the samples stay unchanged. Work on a copy if you edit a sample.

Worked example

import csv
from io import StringIO
source = StringIO("subject,minutes\nPython,20\nSQL,25\n")
rows = []
reader = csv.DictReader(source)
if reader.fieldnames != ["subject", "minutes"]:
    raise ValueError("unexpected CSV columns")
for row in reader:
    line_number = reader.line_num
    if None in row or not isinstance(row["subject"], str) or not row["subject"].strip():
        raise ValueError(f"invalid row on line {line_number}")
    subject = row["subject"].strip()
    try:
        value = int(row["minutes"])
    except (TypeError, ValueError) as error:
        raise ValueError(f"bad minutes on line {line_number}") from error
    if value <= 0:
        raise ValueError(f"invalid row on line {line_number}")
    rows.append((subject, value))
print(rows)

[('Python', 20), ('SQL', 25)]

Why this works

The header is line 1, so the first data row ends on line 2. reader.line_num reports the physical line where the current CSV record ends, including a record with a quoted newline. strip() normalises surrounding subject whitespace, and int() makes numeric operations possible.

Another case

import csv
from io import StringIO
print(next(csv.DictReader(StringIO("subject,minutes\nPython,20\n")))["minutes"])

20

Practice

Try a row with minutes 0. What should the validation report? Then run the same checks on the sample sessions.csv from the practice bundle by opening the file instead of using StringIO; you should get [('Python', 20), ('SQL', 25), ('Python', 30)].

Try another case

In the same CSV, a third line SQL,ten replaces SQL,25. What error appears?