Files, CSV and memory that survives

You will be able to

• ✓ Explain why files exist (memory vs disk)
• ✓ Model a CSV as a list of dicts
• ✓ Know open/read/write and the csv module at a professional level

RAM forgets; disks remember

A variable lives in RAM. Close the laptop, it is gone. A file is bytes on disk with a path: data/attendance.csv. Applications read at startup, write when something important changes, and never assume the previous run is still in memory.

Paths are not the same on every computer. pathlib.Path("data") / "marks.csv" joins folders correctly on Windows and Linux. Hard-coding C:\\Users\\Asha\\Desktop\\file.csv is how student projects die on the teacher’s machine.

CSV is a table in text

First line: column names. Next lines: values in the same order. Commas separate fields; quotes wrap a field that itself contains a comma. That is the whole format. It is not a database — there is no index, no type, no two-user lock — but it is the interchange format of science and government.

After you parse a CSV you almost always want a list of dicts: each row is {"name": "Meera", "present": 1}. Then filtering, grouping and writing back are the same loops you already know. pandas.read_csv does this parse for you; you should be able to picture the list of dicts it produces.

open, with, and encoding

with open(path, encoding="utf-8") as f: is the professional open. with closes the file even if you crash. encoding="utf-8" is how Indian names survive. Never open a text file without saying the encoding — Windows otherwise guesses, and गुजरात becomes garbage.

Write atomically when you can: write to marks.csv.tmp then rename. A crash mid-write should not leave a half CSV as the only copy of the term’s marks.

What not to store in CSV

Passwords, tokens, and anything you would not pin on the staff-room noticeboard. CSV has no encryption. Nested data (a student with a list of achievements) fits JSON or a database better than ten extra columns named badge1, badge2.

Very large files (millions of rows) still start as CSV in the wild. pandas and databases exist because scanning a 2 GB CSV on every page load is too slow — which is the next chapter.

Words that matter

Persistence

Data that still exists after the process exits.

CSV

Comma-separated values: a text table with a header row.

Encoding

How characters become bytes; use UTF-8.

pathlib

Standard-library paths that work across operating systems.

Common mistakes

On a full Python install — csv + pathlib

The standard library already reads and writes CSV correctly, including quoted commas. pandas is optional sugar on top of this shape.

No pip needed.

Read attendance.csv into a list of dicts, then write a filtered file.

import csv
from pathlib import Path

path = Path("data/attendance.csv")
rows = []
with path.open(encoding="utf-8", newline="") as f:
    for row in csv.DictReader(f):
        rows.append(row)

present = [r for r in rows if r["status"] == "P"]

out = Path("data/present.csv")
with out.open("w", encoding="utf-8", newline="") as f:
    w = csv.DictWriter(f, fieldnames=["name", "status"])
    w.writeheader()
    w.writerows(present)

Example program — A CSV already loaded as rows

This is what DictReader gives you. Filter present students.

Your turn — Keep the science rows

rows has subject keys. Collect names where subject is Science. Print so Anil appears.

Self-assessment

Check your understanding before you mark the lesson complete.