File Handling — notes

A file keeps data on disk after a program ends (data in variables are lost).

• Text file: characters in lines; each line ends with the EOL character \n. Readable in any editor.
• Binary file: raw bytes, e.g. Python objects stored with pickle; not human-readable.
• CSV file: a text file of comma-separated values, one record per line; opens in spreadsheets.

Paths. An absolute path starts at the root (C:\School\cs\a.txt, /home/asha/cs/a.txt). A relative path starts at the current working directory (a.txt, data\a.txt, ..\a.txt, where .. is the parent folder).

f = open(name, mode) returns a file object; f.close() saves and releases the file.

• "r" read (default; file must exist), "w" write (creates or empties), "a" append (creates or adds at the end).
• "r+", "w+", "a+" add the other operation; add b for binary: "rb", "wb", "ab".

The with statement closes the file automatically, even after an error:

with open("notes.txt", "a") as f:
    f.write("New entry\n")

Writing: write(s) writes one string and returns its length; writelines(list) writes each string. Neither adds \n. Convert numbers with str().

Reading: read() → whole remaining text as one string; read(n) → n characters; readline() → one line ("" at the end of the file); readlines() → list of lines; for line in f: → one line per pass.

Output of the writing example (run in Python):

Line one
Line two
Line three

tell() gives the file pointer's position in bytes; seek(n) moves it to byte n (seek(0) = start).

Typical processing tasks: count lines, words (split()), characters (len()), or lines that start with a letter; display long words; copy selected lines to another file opened in "w" mode.

with open("routes.txt") as src, open("caps.txt", "w") as dst:
    for line in src:
        if line[0].isupper():
            dst.write(line)

Resulting caps.txt (run in Python):

Route 7 Kothrud
Route 3 Hadapsar
Route 21 Wakad

import pickle. Pickling (pickle.dump(obj, f)) writes an object as bytes; unpickling (pickle.load(f)) reads the next object back.

• Open in "wb" (new file), "ab" (add records) or "rb" (read).
• Read till the end: call load() in while True inside try, and stop on except EOFError.
• Search: load each record and test it; use a flag to report "not found".
• Update: load all records into a list, change in memory, dump all again in "wb".

Appending a fifth record and counting again (run in Python):

import pickle
with open("books.dat", "ab") as f:
    pickle.dump([105, "Blue Kettle", 210], f)
n = 0
with open("books.dat", "rb") as f:
    try:
        while True:
            pickle.load(f)
            n += 1
    except EOFError:
        pass
print(n)

Output: 5

import csv. Open with newline="" so that the writer controls line endings.

• Writing: w = csv.writer(f), then w.writerow(list) or w.writerows(list_of_lists).
• Reading: for row in csv.reader(f): gives each line as a list of strings; next(reader) skips the header; convert numbers with int().

import csv
with open("marks.csv", newline="") as f:
    for row in csv.reader(f):
        print(row)

Output (run in Python):

['roll', 'name', 'marks']
['1', 'Aarav', '78']
['2', 'Bina', '91']
['3', 'Chetan', '64']
['4', 'Divya', '85']