A document from MCS 275 Spring 2023, instructor David Dumas. You can also get the notebook file.

MCS 275 Spring 2023 Worksheet 12 Solutions

  • Course instructor: David Dumas
  • Solutions prepared by: Johnny Joyce, Jennifer Vaccaro, Kylash Viswanathan, David Dumas

Topics

This worksheet focuses on SQLite databases. (Last week we also had one lecture about HTML, and we'll come back to that in the next worksheet.)

Resources

These things might be helpful while working on the problems. Remember that for worksheets, we don't strictly limit what resources you can consult, so these are only suggestions.

The main references for these topics are:

Work on these three setup steps before lab

If possible. But if you can't, you can work on this at the start of lab.

Step 1: Get the SQLite command line shell

There's nothing to install if you just want to use SQLite with Python, as the module sqlite3 is already part of the Python standard library.

However, if you just want to run one SQL query and show the results, a fairly long Python program is needed, and the results are not presented in a convenient form for experimentation (column headings etc.).

It's much easier if you can install the SQLite REPL (or "command line shell") that lets you run queries directly from a prompt in your terminal, so we strongly recommend installing the command line shell. The instructions in the rest of this worksheet assume you have the command line shell, and you'll need to translate them into corresponding Python programs if you want to avoid installing it.

Lectures 28 and 29 discussed how to install the command line shell. For MacOS and Linux, there's usually nothing you need to do. Here is a quick video showing the steps to install it under windows 10. Other recent Windows versions will be similar.

Here are more detailed written installation instructions by platform.

  • If you use Windows, you need to install it yourself. The installation doesn't look like the graphical ones you may be used to, with a window and buttons guiding you through the steps. Instead you download a zip file and extract it. The whole thing is relatively quick, but the steps below are described in some detail, so the written instructions are a bit long. (Consider just watching the 1-minute video and following along instead.)

    1. Go to https://www.sqlite.org/download.html.

    2. Scroll down to the heading Precompiled Binaries for Windows

    3. In that section, click on the link whose name begins with sqlite-tools-win32-x86-. The description next to the link should begin: A bundle of command-line tools for managing SQLite database files

    4. The download should either begin immediately, or you'll be asked if you want to download the file. If prompted, save it somewhere you know how to find in the file explorer.

    5. Find the ZIP file you just downloaded and open it in the file explorer (either by clicking the filename where the web browser shows a recent download, or by opening the file explorer and navigating to the place where you saved the ZIP file).

    6. This should open a new explorer window in which you browse the contents of the ZIP file. It contains a folder with a name like sqlite-tools-win32-x86-3380200. Double click to enter that folder.

    7. You should now see a list of three files, named

      • sqldiff.exe
      • sqlite3.exe
      • sqlite3_analyzer.exe

      However, the .exe may be missing if explorer is configured to hide file extensions (the default).

    8. Of these, sqlite3.exe is the only one you want. Drag that file to the desktop to extract a copy of it.

    9. You should now see a sqlite3 icon on the desktop. Don't click it; we'll work in the terminal instead.

    10. Open a terminal (Windows Terminal or PowerShell) and change directory to the desktop (e.g. with cd Desktop or cd C:\Users\myusername\Desktop or cd C:\Users\myusername\OneDrive\Desktop, depding on where PowerShell opens and whether you use OneDrive desktop backup)

    11. Test running the command line shell by entering this command in powershell: .\sqlite3.exe

    12. Now SQLite should display a startup message and prompt. Success!

    13. Exit the SQLite command line shell with .exit to return to PowerShell

    14. In the future, you will probably want to run the SQLite command line shell from whatever directory contains your database file. To do that, open PowerShell, go to the directory you want to work in, and then enter the full path to the sqlite3.exe file, e.g. C:\Users\myusername\Desktop\sqlite3.exe.

      • If your username contains spaces, this won't work. You will need to instead surround the command with single quotation marks and precede it with an ampersand and space, e.g.
      & 'C:\Users\My Username Has Spaces\Desktop\sqlite3.exe'

  • If you use Linux or MacOS, SQLite's command line shell is almost always pre-installed. Type sqlite3 in a terminal and press enter. Success (meaning it is already installed) looks something like this:

sqlite3
SQLite version 3.31.1 2020-01-27 19:55:54
Enter ".help" for usage hints.
Connected to a transient in-memory database.
Use ".open FILENAME" to reopen on a persistent database.
sqlite>

(at which point you'd want to exit using command .exit) Failure (meaning it is not already installed) looks something like this:

$ sqlite3
sqlite3: command not found

If you use Linux or MacOS and SQLite's command line shell is not already installed, contact your TA or instructor for help.

Step 2: Know how to tell which REPL is open, and how to exit SQLite

Any time you have your terminal open, there are three situations you may find yourself in:

  1. The terminal itself is waiting for a command from you. If this is the case, you'll see a prompt like
PS C:\Users\ddumas\Desktop>

or

$
  1. The Python REPL is running, and waiting for a Python statement from you. If this is the case, you'll see a prompt
>>>

and you can quit back to the terminal using the command exit().

  1. The SQLite command line shell is running, and is waiting for a SQL command from you. If this is the case, you'll see a prompt
sqlite>

and you can quit back to the terminal using the command .exit

To summarize

When running prompt looks like exit with and then
Terminal (Windows) PS C:\Users> exit window closes
Terminal (Mac/Linux) $ exit window closes
Python >>> exit() back to terminal
SQLite sqlite> .exit or .quit back to terminal

Step 3: Get the sample database and confirm you can open it

So you'll need to download this and put the file in a place where you can find it.

The only table in this database is called powerplants, and the columns present in that table are described at

To confirm you can open the database:

  • Open your terminal
  • Change directory to the one that contains the file powerplants.sqlite
  • Run the SQLite command line shell, giving powerplants.sqlite as the first command line argument, e.g. a typical Windows command for PowerShell would be
    C:\Users\myusername\Desktop\sqlite3.exe powerplants.sqlite
    and a typical Linux or MacOS terminal command would be
    sqlite3 powerplants.sqlite
    After pressing enter, you should see that SQLite is running and waiting for a command with the prompt
    sqlite>
  • KEY STEP: We need to check that SQLite actually opened the star database file, because if SQLite didn't find it, then the command we ran will actually create an empty file with that name. To check that we actually opened an existing database, run the SQLite command line shell command to list tables in the database:
    .tables
    If everything is working, the output should look like this:
    sqlite> .tables
powerplants
sqlite>
    If you instead see this:
    sqlite> .tables
sqlite>
    then it means you ran sqlite3 successfully, but in a directory that didn't contain the star database. That directory will now contain an empty file named powerplants.sqlite, which you should probably find and remove to prevent yourself from later confusing it with the actual database file.

1. Power plant queries

I recommend working on these in the SQLite command line shell. Remember, the commands

.headers on
.mode column

will tell SQLite to present output in a more readable format.

Use SQL queries run against the power plant database to answer these questions. (You should consider both the query and its return value as part of the answer.)

1A. Largest by country

For each country below, what is the name, primary fuel, and generation capacity (in MW) of the largest power plant in the database?

  • United States of America
  • India
  • Brazil
  • China
  • Australia
  • France

1B. Hyperborean radionuclides

Find the northernmost nuclear power plant in the database. What country is it in? What is its location (in latitude, longitude)? What is its generation capacity?

1C. Almost prime meridian wind

What countries contain wind power plants listed in this database that are located within 5 degrees longitude on either side of the prime meridian?

And among these, which ones contain large wind plants, if we define that as 200MW or greater?

1D. Output availability

There are columns in the database for net energy output in each of the years 2014 to 2019. But in many cases this information is not available.

What fraction of the power plants listed in the database have output information for 2017?

What fraction of the power plants listed in the database have output information for all of the years 2014 to 2019?

Solution:

1A

Note that the country parameter takes on each of the values shown in question 1A.

SELECT country, name, primary_fuel, capacity_mw
FROM powerplants 
WHERE country = "United States of America"
ORDER BY capacity_mw DESC
LIMIT 1;


country                   name          primary_fuel  capacity_mw
------------------------  ------------  ------------  -----------
United States of America  Grand Coulee  Hydro         6809.0

Repeating similarly for the other countries we can assemble this table:

country                   name               primary_fuel  capacity_mw
------------------------  ------------       ------------  ----------------
United States of America  Grand Coulee       Hydro         6809.0  
India                     VINDH_CHAL STPS    Coal          4760.0   
Brazil                    Tucuruí            Hydro         8535.0  
China                     Three Gorges Dam   Hydro         22500.0  
Australia                 Bayswater          Coal          2640.0 
France                    GRAVELINES         Nuclear       5460.0

1B

SELECT name, country, latitude, longitude, capacity_mw 
FROM powerplants 
WHERE primary_fuel = "Nuclear"
ORDER BY latitude DESC
LIMIT 1;


name      country  latitude  longitude  capacity_mw
--------  -------  --------  ---------  -----------
Bilibino  Russia   68.0503   166.5389   48.0

1C

SELECT DISTINCT country 
FROM powerplants 
WHERE primary_fuel = "Wind" AND longitude <= 5 AND longitude >= -5;

country       
--------------
Belgium       
France        
Morocco       
Netherlands   
Spain         
United Kingdom

Countries containing wind power plants within 5 degrees of the prime meridian, that have capacity of at least 200 mW

SELECT DISTINCT country 
FROM powerplants 
WHERE primary_fuel = "Wind" AND longitude <= 5 AND longitude >= -5 AND capacity_mw >= 200;


country       
--------------
Belgium       
Netherlands   
United Kingdom

1D

The following query extracts a count of those columns in which the attribute 'output_gwh_2017' is not missing

SELECT count(*) 
FROM powerplants 
WHERE output_gwh_2017 IS NOT NULL;

count(*)
--------
9500

The following query obtains a total count of the rows

SELECT count(*)
FROM powerplants;

count(*)
--------
34936
In [3]:
data_Avail = 9500
total_count = 34956
'The percentage of plants with output data in 2017 is {:2.2%}'.format(9500/34956)
Out[3]:
'The percentage of plants with output data in 2017 is 27.18%'

We can similarly compute the results of the query for all years from 2014 thru 2017 that contain data, but logically constraining the attributes with an AND operator

SELECT count(*)
FROM powerplants
WHERE (output_gwh_2014 IS NOT NULL) AND (output_gwh_2015 IS NOT NULL) AND (output_gwh_2016 IS NOT NULL)
AND (output_gwh_2017 IS NOT NULL) AND (output_gwh_2018 IS NOT NULL) AND (output_gwh_2019 IS NOT NULL);

count(*)
--------
6483

The result of this query is 6483, meaning the ratio is given by:

In [4]:
data_Avail = 6483
total_count = 34956
'The percentage of plants with output data in from 2014-2019 is {:2.2%}'.format(6483/34956)
Out[4]:
'The percentage of plants with output data in from 2014-2019 is 18.55%'

2. Programming joke utility

Expressions in SQL

There is a new SQLite feature (or more precisely, a feature of the SQL dialect that SQLite uses for queries) that you'll need to use in this problem.

In many places where we've used column names in our queries, you can also use expressions that apply arithmetic operators and other functions to the values in the columns. For example, if a database of MCS 275 grades has columns called project3pct and project4pct, then this query would return the email addresses of students whose grades on those two projects differed by more than 10 percent:

SELECT email FROM mcs275roster WHERE ABS(project3pct - project4pct) > 10;

You can also use expressions like this in the requested list of output columns. For example, this query would get the average of project 3 and 4 percentages for all students, listed in alphabetical order by last name.

SELECT lastname, firstname, 0.5*(project3pct + project4pct) FROM mcs275roster ORDER BY lastname;

Such expressions can also be used after ORDER BY to make a custom sort.

You can find lists of built-in functions in SQLite in the documentation:

The task

Write a program that stores, delivers, and ranks programming jokes using a SQLite database. It should support three operations:

  • Add a joke: The user provides a string and the program adds it to the joke table. In addition to the joke itself, the joke table has columns to store the number of times a joke has been told, and how many times it was successful in amusing the user.
  • Tell a joke: A random joke is selected and displayed. The user is asked if they found it amusing. Their answer is recorded by updating the relevant counts (times told, times successful) for that row.
  • Best jokes list: Among jokes that have succeeded at least once, display the top 5 ranked by success rate (ratio of times successful to times told).

The program should create the database and table it needs if they don't already exist. Otherwise, it should open and use the existing database.

The three functions should be selected using command line arguments. The first command line argument is always the command---one of add, tell, or best. If the command is add, then a second command line argument is required, which is the joke itself. If the command is tell, no other arguments are required but the user is prompted for their approval/disapproval of the joke through keyboard input.

Hints:

  1. Between lectures 29 and 30, I added a sample "todo list" application to the course sample code repository. Taking a look at that might be a good way to start.
  2. I suggest making a table that has a column of type INTEGER PRIMARY KEY so you have a way to uniquely refer to rows.
  3. SQLite doesn't have a feature to select a random row directly, but it does have a built-in function RANDOM() that will return a random number for each row, and you can order by that. Ordering by RANDOM() and limiting the number of rows to 1 is equivalent to making a random choice.
  4. In a SQLite query, if columns A and B both have type integer then A/B computes the integer division of A by B. In contrast, 1.0*A/B would give the true quotient because the multiplication by 1.0 converts A to a float (or REAL, in SQLite terminology).

Save the program as jokedb.py.

Here is a sample session of what using it should look like. (These are from a linux terminal session, where the terminal prompt is "$". In Windows PowerShell, the prompt will look a bit different.)

$ python3 jokedb.py tell
ERROR: No jokes in database.

[... now suppose several jokes are added ...]

$ python3 jokedb.py tell
There are 10 types of people in the world: Those who understand binary, and those who don't.

Were you amused by this? (Y/N)y

$ python3 jokedb.py tell
The two hardest things in programming are naming things, cache invalidation, and off-by-one errors.

Were you amused by this? (Y/N)n

$ python3 jokedb.py add "Most people agree that there were no widely-used high-level programming languages before FORTRAN.  Unfortunately, there is no agreement on whether this makes FORTRAN the 1st such language, or the 0th."

$ python3 jokedb.py tell
Python developers rarely decorate the walls of their offices. They consider whitespace to be significant.

Were you amused by this? (Y/N)y

$ python3 jokedb.py best
-------------------------------------------------------
#1 with 100% success rate after 8 tellings:
Knock knock.
Race condition.
Who's there?
-------------------------------------------------------
#2 with 71% success rate after 7 tellings:
There are 10 types of people in the world: Those who understand binary, and those who don't.
-------------------------------------------------------
#3 with 67% success rate after 6 tellings:
A software testing engineer walks into a bar and orders a refrigerator, -1 glasses of water, and INT_MAX+1 cans of soda.
-------------------------------------------------------
#4 with 60% success rate after 5 tellings:
Python developers rarely decorate the walls of their offices. They consider whitespace to be significant.
-------------------------------------------------------
#5 with 50% success rate after 4 tellings:
The two hardest things in programming are naming things, cache invalidation, and off-by-one errors.

$

Solution

In [ ]:
import sys
import sqlite3

def add(con, joke):
    """If db table exists, add a joke. Otherwise, create a table and add the joke."""
    
    # If the table does not exist yet, create it
    con.execute("CREATE TABLE IF NOT EXISTS jokes (joke TEXT, tells INT,success INT)")
    
    # By now, table must exist. 
    # We haven't told the joke yet, so other vals set to 0
    con.execute("INSERT INTO jokes VALUES (?,?,?);",(joke,0,0))
    
    con.commit()

def tell(con):
    """Tells a random joke from the database"""
    
    # Select a random joke and its corresponding row data
    c = con.execute("SELECT rowid,* FROM jokes ORDER BY RANDOM() LIMIT 1")
    
    # Unpack the values and print out the joke
    rowid, joke, tells, success = c.fetchone()
    print(joke)
    tells += 1
    
    # While loop to account for invalid user input
    while True: 
        rank = input("Was this a good joke? (Y/N)")
        if rank.upper() == "Y":
            # It's a good joke!
            success += 1
            break
        elif rank.upper() == "N":
            # It's a bad joke :-(
            success += 0
            break
        else:
            print("Invalid input, please try again.")
            
    # Update the tells and success values for the joke we told (identified by rowid)
    con.execute("UPDATE jokes SET tells=?,success=? WHERE rowid==?;",(tells,success,rowid))
    con.commit()

def best(con):
    """Prints out the top 5 jokes, based on ratings"""
    c = con.execute("SELECT * FROM jokes WHERE tells>0 ORDER BY -1.0*success/tells LIMIT 5")
    for i,joke in enumerate(c):
        best_print(i+1,joke[0],joke[1],joke[2])

def best_print(num, joke, tells, successes):
    """Prints joke data with a specific format"""
    print("-------------------------")
    if tells == 0:
        print("#{} with no tellings yet:")
    else:
        print("#{} with {}% success rate after {} tellings:".format(num,int(100*successes/tells),tells,joke))
    print(joke)

if __name__=="__main__":
    cmd = sys.argv[1]
    con = sqlite3.connect("jokes.sqlite")

    # Check the command value, and call the appropriate function
    if cmd == "add":
        joke = sys.argv[2]
        add(con,joke)
    elif cmd == "tell":
        tell(con)
    elif cmd == "best":
        best(con)
    else:
        print("Command not understood. Please use 'add', 'tell', or 'best'.")
    
    # Close the connection (any changes were commited within the function)
    con.close()

3. Extra: Power plant visualization

Work on this if you complete the problems above.

Here's a scatter plot of all the powerplants in the database that can produce at least 200MW, and the matplotlib code that created it.

In [28]:
import os
os.chdir("/home/ddumas/teaching/mcs275/public/samplecode/sql")
In [29]:
import matplotlib.pyplot as plt
import numpy as np
import sqlite3

con = sqlite3.connect("powerplants.sqlite")
res = con.execute("""
SELECT latitude,longitude
FROM powerplants
WHERE capacity_mw > 200;
""")

longitudes = []
latitudes = []
for row in res:
    latitudes.append(row[0])
    longitudes.append(row[1])

# Optional: Convert to arrays, which would make it easier
# to do arithmetic or transformations on them.
longitudes = np.array(longitudes)
latitudes = np.array(latitudes)

# Make the plot
plt.figure(figsize=(8,4),dpi=120)
plt.scatter(longitudes,latitudes,s=0.5)
plt.xlabel("longitude")
plt.ylabel("latitude")
plt.show()

Refine this to plot as follows:

  • Only show power plants in the lower 48 states of the USA (limit by country in query and adjust axes limits in matplotlib)
  • Have different fuels shown in different colors
  • Make the size of each dot depend on the capacity of the powerplant

(For a real geographic visualization, you'd want to do more---such as adding a legend describing which colors correspond to which types, the outlines of the continents, etc.---but this is a start!)

In [1]:
import matplotlib.pyplot as plt
import numpy as np
import sqlite3

con = sqlite3.connect("powerplants.sqlite")

res = con.execute("""
SELECT DISTINCT primary_fuel FROM powerplants
WHERE capacity_mw > 200
AND country = "United States of America";
""")
fuels = [x[0] for x in res.fetchall()]

res = con.execute("""
SELECT latitude,longitude,primary_fuel,capacity_mw
FROM powerplants
WHERE capacity_mw > 200
AND country = "United States of America";
""")

longitudes = []
latitudes = []
kinds = []
sizes = []
for row in res:
    latitudes.append(row[0])
    longitudes.append(row[1])
    kinds.append(fuels.index(row[2]))
    sizes.append(row[3])

# Optional: Convert to arrays, which would make it easier
# to do arithmetic or transformations on them.
longitudes = np.array(longitudes)
latitudes = np.array(latitudes)
kinds = np.array(kinds)
sizes = np.array(sizes)

# Make the plot
plt.figure(figsize=(6.5,4.5),dpi=100)
plt.scatter(longitudes,latitudes,s=0.01*sizes,c=kinds,alpha=0.5)
plt.xlabel("longitude")
plt.ylabel("latitude")
plt.xlim(-130,-65)
plt.ylim(20,55)
plt.show()