Final Exam Model Paper 3

i. Inside the function, discount is 25, so the price is 400 − 400 × 25 // 100 = 400 − 100 = 300. Outside, discount is still 10:

300 10

ii. Line 4 makes a new local variable called discount inside the function. It shadows the global one: inside the function, Python finds the local name first. The global discount is never changed, and the local one disappears when the function returns.

• i.Any two: the lanes’ shapes and bends, their widths, what they are made of, the shops and houses along them, their lengths. This is abstraction: the graph keeps only what the question needs, which corners are joined to which, as Euler kept only which land each bridge joined. Everything left out makes no difference to which corners can be reached, or in how many steps.
• ii.Each edge needs a weight, the lane’s length in meters: it becomes a weighted graph. Without weights every lane counts as one step, so a search finds the fewest lanes, which may be a longer walk than a route with more, shorter lanes.

CREATE TABLE shops (  license_no INTEGER PRIMARY KEY,  name TEXT NOT NULL,  ward INTEGER CHECK (ward BETWEEN 1 AND 9),  phone TEXT UNIQUE);

The phone number is TEXT, not a number: nobody adds phone numbers, and as a number a leading 0 or a + would be lost.

import pandas as pd scores = pd.read_csv("teej.csv")means = scores.groupby("group")["score"].mean()print(means.sort_values(ascending=False))

groupBhimsen Tole      84.0Thado Tole        80.0Pipal Chautari    72.0Name: score, dtype: float64

Bhimsen Tole: (85 + 88 + 79) ÷ 3 = 84. Thado Tole: (78 + 82 + 80) ÷ 3 = 80. Pipal Chautari: (70 + 74 + 72) ÷ 3 = 72.

• i.Real companies don’t phone you about a virus, and they don’t advertise support numbers in pop-ups; fake “support lines” pay for pop-ups and adverts like this one. The program he is asked to install would let a stranger control his screen, and with it his files, accounts and money apps; next the man would ask for passwords, codes or a payment.
• ii.Not install anything and hang up; close the pop-up (or the browser) without calling, and never give a password or a code sent to his phone. Then fix the slow laptop the usual way, looking first at what is using the disk and memory, or ask a teacher or someone he trusts.

i. marks.csv is one level up from project, so:

cat ../marks.csv

ii. An absolute path starts from the very top, /, and Bina’s home folder is /home/bina:

cat /home/bina/school/class9/notes.txt

(cat ~/school/class9/notes.txt names the same file from the home folder.)

• i.Chatbots are trained on people’s ratings of their answers, and people tend to rate answers that agree with them more highly. So a model can learn to agree too easily, even with a wrong idea. It has no way to look the fact up; it predicts the reply that sounds most welcome after “I think it is Birgunj”.
• ii.Ask a neutral question, without saying what he thinks the answer is, and don’t treat its answer as checked: check the fact in a reliable source, such as the province government’s website or a textbook.

His check is not enough. One test of an easy case can’t show that the code does what the offer says. Tests should come from the requirement, including its edge: “Rs 1,000 or more” means a bill of exactly Rs 1,000 gets 20% off. He should test 999 (no discount), 1000 (800) and 1001, and one ordinary value. holi_bill(1000) gives 1000, so the test finds the mistake: the AI wrote > where the offer needs >=. He must understand the code he uses, not just see it run once.

Any two: fix the dates from more than one source, and say where sources disagree and what each counts; prefer sources from the time (a newspaper report of the day, a law’s own text) to later summaries; ask what made each step possible, and what held it back (money, laws, needs, the land, disasters, other technologies); ask who gained and who was left out at each step; ask what the technology did to society and the environment, intended or not.

The claim goes far beyond its evidence. First, the Kenyan figure comes from one study whose methods other researchers have seriously questioned; economists still debate how much M-Pesa reduced poverty. Second, even taken at face value, it is 2 households in 100, not the end of poverty. Third, Nepal isn’t Kenya: M-Pesa worked where most people had no bank account and families sent money home by hand. And even in Kenya some were left out: people with no phone, far from any agent, or whose savings were eaten by fees. A fair claim would say what mobile money might help with, for whom, and how we would find out.

i. and ii.

"""Change for the shop's customers.""" NOTES = [1000, 500, 100, 50, 20, 10, 5, 1]    # rupees, the biggest first  def change_notes(paid, price):    """The change, as {note: how many}, in as few notes as possible.    Raises ValueError if paid is less than price."""    if paid < price:        raise ValueError("paid less than the price")    left = paid - price    change = {}    for note in NOTES:        if left >= note:            change[note] = left // note            left = left % note    return change  if __name__ == "__main__":    print(change_notes(500, 500), "should be {}")    print(change_notes(100, 37), "should be {50: 1, 10: 1, 1: 3}")    try:        change_notes(100, 120)        print("no error: wrong")    except ValueError:        print("ValueError, as it should be")

iii.

import change for note, count in change.change_notes(2000, 765).items():    print("Rs", note, "×", count)

Rs 1000 × 1Rs 100 × 2Rs 20 × 1Rs 10 × 1Rs 5 × 1

The change is 2,000 − 765 = Rs 1,235: one 1,000, then 235 left: two 100s, then 35: one 20, one 10, one 5. Taking the biggest note that fits each time gives the fewest notes for Nepal’s notes. Any clear names are fine.

i. Each path goes both ways, so it is written in both places’ lists:

ponds = {    "Janaki Mandir": ["Gangasagar", "Dhanushsagar"],    "Gangasagar": ["Janaki Mandir", "Dhanushsagar", "Ratnasagar"],    "Dhanushsagar": ["Janaki Mandir", "Gangasagar", "Argaja"],    "Ratnasagar": ["Gangasagar", "Argaja", "Police post"],    "Argaja": ["Dhanushsagar", "Ratnasagar", "Bihari Kund"],    "Bihari Kund": ["Argaja"],    "Police post": ["Ratnasagar"],}

• ii.Janaki Mandir 2, Gangasagar 3, Dhanushsagar 3, Ratnasagar 3, Argaja 3, Bihari Kund 1, Police post 1. Together 16 = 2 × 8 paths, because each path has two ends, so it is counted once in each of the two places’ lists.
• iii.1 path away: Gangasagar, Dhanushsagar. 2 paths away: Ratnasagar (from Gangasagar), Argaja (from Dhanushsagar). 3 paths away: Police post (from Ratnasagar), Bihari Kund (from Argaja).

• Bihari Kund: 3 paths, Janaki Mandir → Dhanushsagar → Argaja → Bihari Kund.
• Police post: 3 paths, Janaki Mandir → Gangasagar → Ratnasagar → Police post.

• i.The names that start with “Abir” and are in stock: Abir red, packet and Abir green, packet. The pink has qty 0.
• ii.Prices from 100 to 200, both ends included, the cheapest first:

• iii.With no WHERE, the UPDATE changed every row: every item now costs Rs 500. He should have typed UPDATE stock SET price = 500 WHERE item_id = 4;.
• iv.The three packets of abir cost less than Rs 100: 3 rows, with 120 + 80 + 0 = 200 packets in all.

• i.A random number from the normal distribution (the bell shape) with its middle at 100 and a standard deviation of 15: most days near 100 cups, about two in three between 85 and 115, and a few much busier or quieter.
• ii.It works out the liters of milk that day’s cups need (1 liter for 8 cups) and checks whether that is more than 15 liters. 15 × 8 = 120, so a day of more than 120 cups is a busy day.
• iii.One run of 30 days is one possible month; another month could have more or fewer busy days just by chance. Running it many times and looking at the results together (their average and their spread) shows what is usual, and how much it can vary.
• iv.Compare it with real records: count the cups she really sold on past days, and see whether their middle and spread look like 100 and 15 (and whether busy days really are as common as the simulation says). The model also leaves things out, such as festivals, market days and the tourist season, that she could check for.

• i.A part whose failure stops the whole system. The first time it was the one server: when it stopped, nobody could buy.
• ii.It stands in front of several servers doing the same job and passes each request to one of them, sharing out the work.
• iii.The health check is the load balancer asking each server every few seconds whether it is working; failover is switching visitors to the working servers when one fails.
• iv.Scaling is growing or shrinking a system to fit the work. In the cloud, servers are rented and can be added in minutes for the busy hours and removed afterwards, paying only for what is used, instead of buying enough servers for the busiest hour and leaving them idle.

"""Categories for the half marathon."""  def category(age):    """The category's name for a runner of this age, or "Too young"."""    if age < 10:        return "Too young"    if age < 16:        return "Junior"    if age < 40:        return "Open"    if age < 60:        return "Veterans"    return "Seniors"  if __name__ == "__main__":    # tests on both sides of every edge    for age, should in [(9, "Too young"), (10, "Junior"), (15, "Junior"),                        (16, "Open"), (39, "Open"), (40, "Veterans"),                        (59, "Veterans"), (60, "Seniors")]:        print(age, category(age), "should be", should)

Run by itself, race.py prints its tests:

9 Too young should be Too young10 Junior should be Junior15 Junior should be Junior16 Open should be Open39 Open should be Open40 Veterans should be Veterans59 Veterans should be Veterans60 Seniors should be Seniors

import race for runner in range(1, 5):    age = int(input(f"Runner {runner}: age? "))    print("Category:", race.category(age))

Runner 1: age? 15Category: JuniorRunner 2: age? 40Category: VeteransRunner 3: age? 72Category: SeniorsRunner 4: age? 8Category: Too young

The same program, with these three lines in place of the last one:

import sqlite3 db = sqlite3.connect(":memory:")db.execute("CREATE TABLE services (service_id INTEGER PRIMARY KEY, name TEXT, fee INTEGER)")db.execute("""CREATE TABLE requests (request_id INTEGER PRIMARY KEY,              service_id INTEGER REFERENCES services(service_id),              day TEXT, status TEXT)""")db.executemany("INSERT INTO services VALUES (?, ?, ?)", [    (1, "Recommendation letter", 200),    (2, "Birth registration", 0),    (3, "Relationship certificate", 300),    (4, "House map approval", 1500),    (5, "Business registration", 500),])db.executemany("INSERT INTO requests VALUES (?, ?, ?, ?)", [    (1, 1, "2026-09-01", "done"),    (2, 2, "2026-09-01", "done"),    (3, 1, "2026-09-02", "waiting"),    (4, 5, "2026-09-02", "done"),    (5, 1, "2026-09-03", "done"),    (6, 3, "2026-09-03", "waiting"),    (7, 2, "2026-09-04", "done"),    (8, 1, "2026-09-04", "done"),    (9, 5, "2026-09-05", "waiting"),    (10, 5, "2026-09-05", "done"),])  def ask(sql):    """Print each row the query gives, then a blank line."""    for row in db.execute(sql):        print(row)    print()  ask("""SELECT request_id, day FROM requests       WHERE status = 'waiting' AND day BETWEEN '2026-09-02' AND '2026-09-04'       ORDER BY day""")ask("SELECT service_id, COUNT(*) FROM requests GROUP BY service_id ORDER BY COUNT(*) DESC")ask("""SELECT services.name, COUNT(requests.request_id) FROM services       LEFT JOIN requests ON requests.service_id = services.service_id       GROUP BY services.service_id""")

(3, '2026-09-02')(6, '2026-09-03') (1, 4)(5, 3)(2, 2)(3, 1) ('Recommendation letter', 4)('Birth registration', 2)('Relationship certificate', 1)('House map approval', 0)('Business registration', 3)

(The requests are made up for this paper.)

site = {    "Home": ["Notices", "Services", "Contact"],    "Notices": ["Home", "Budget"],    "Services": ["Tax", "Ward offices"],    "Tax": ["Services", "Budget"],    "Ward offices": ["Ward 4"],    "Ward 4": ["Contact"],    "Contact": [],    "Budget": ["Notices"],    "Flood notice": ["Home"],}  def fewest_steps(graph, start, goal):    """The route from start to goal with the fewest steps, or None."""    came_from = {start: None}    to_visit = [start]    while len(to_visit) > 0:        page = to_visit.pop(0)        if page == goal:            route = []            while page is not None:                route.append(page)                page = came_from[page]            route.reverse()            return route        for nxt in graph[page]:            if nxt not in came_from:                came_from[nxt] = page                to_visit.append(nxt)    return None  for start, goal in [("Home", "Ward 4"), ("Tax", "Home"), ("Home", "Flood notice")]:    route = fewest_steps(site, start, goal)    if route is None:        print(start, "to", goal + ": Can't get there")    else:        print(start, "to", goal + ":", " → ".join(route), "(" + str(len(route) - 1), "clicks)")

Home to Ward 4: Home → Services → Ward offices → Ward 4 (3 clicks)Tax to Home: Tax → Budget → Notices → Home (3 clicks)Home to Flood notice: Can't get there

Last monsoon’s Flood notice links to Home, but no page links to it any more, so it can’t be reached by clicking.

• i.A returned value can be used by the program that calls it: main.py prints it here, but another program could count the runners in each category or write them in a file. A function that printed would decide that for every user. 39 and 40 are the two sides of an edge between Open and Veterans, where a <= written for a < would put a runner in the wrong category.
• ii.The LEFT JOIN keeps every service, even with no matching request, and fills the request’s fields with NULL; COUNT(requests.request_id) counts only the fields that aren’t NULL, so it gives 0. A plain JOIN would drop House map approval altogether, since no request matches it; it wouldn’t appear in the list at all.
• iii.The links are one-way: the graph is directed. Home → Services → Tax uses links that exist only in that direction; to go back, Tax must use its own links, Tax → Budget → Notices → Home. On a map of two-way roads (an undirected graph), every route can be walked backwards, so the fewest steps are the same both ways.