Errors.md

Errors

This file presents a diagnostic challenge. Each example includes a buggy program and one or more corrected programs. A good programming media should help programmers to avoid writing these buggy programs or to recover from the bug and finally reach a corrected program.

These examples are adapted from student code collected in CS111 at Brown University.

To keep the authenticity of some error cases, we assume the existence of two plotting functions:

• scatterPlot :: t:Table * c1:ColName * c2:ColName -> Image, where both input columns must contain numbers.
• pieChart :: t:Table * c1:ColName * c2:ColName -> Image, where the first column must contain categorical values, and the second column must contain positive numbers.

Malformed Tables

This section lists errors that programmers can make when constructing table constants. All these malformed tables should be corrected to the students table, which is shown below with a full schema declaration.

| name    | age    | favorite color |
| String  | Number | String         |
| ------- | ------ | -------------- |
| "Bob"   | 12     | "blue"         |
| "Alice" | 17     | "green"        |
| "Eve"   | 13     | "red"          |

missingSchema

This malformed table misses the schema.

| "Bob"   | 12 | "blue"  |
| "Alice" | 17 | "green" |
| "Eve"   | 13 | "red"   |

missingRow

This malformed table misses the content of the last row. (Note: the last row is not a row with 3 missing values, but rather a row with no value.)

| name    | age    | favorite color |
| String  | Number | String         |
| ------- | ------ | -------------- |
| "Bob"   | 12     | "blue"         |
| "Alice" | 17     | "green"        |
|                                   |

missingCell

The first row of this malformed table misses a cell.

| name    | age    | favorite color |
| String  | Number | String         |
| ------- | ------ | -------------- |
| "Bob"   | "blue" |
| "Alice" | 17     | "green"        |
| "Eve"   | 13     | "red"          |

swappedColumns

The rows disagree with the schema on the ordering of the first two columns.

| name   | age     | favorite color |
| String | Number  | String         |
| ------ | ------- | -------------- |
| 12     | "Bob"   | "blue"         |
| 17     | "Alice" | "green"        |
| 13     | "Eve"   | "red"          |

schemaTooShort

The schema specifies that there are two columns. But the rows have three columns.

| name    | age    |
| String  | Number |
| ------- | ------ |
| "Bob"   | 12     | "blue"         |
| "Alice" | 17     | "green"        |
| "Eve"   | 13     | "red"          |

schemaTooLong

The schema specifies that there are four columns. But the rows have three columns.

| name    | age    | favorite number | favorite color |
| String  | Number | Number          | String         |
| ------- | ------ | --------------- |----------------|
| "Bob"   | 12     | "blue"          |
| "Alice" | 17     | "green"         |
| "Eve"   | 13     | "red"           |

Using Tables

This section lists errors in using tables. Each example comes with a context, which lists the used tables, and a task, which states how the table(s) should be used.

midFinal

Context

gradebook

Task

The programmer was asked to visualize as a scatter plot the connection between midterm and final exam grades.

A Buggy Program

> scatterPlot(gradebook, "mid", "final")

What is the Bug?

The "mid" is not a valid column name of gradebook. However, the table contains a "midterm" column.

A Corrected Program

> scatterPlot(gradebook, "midterm", "final")

blackAndWhite

Context

jellyAnon

Task

The programmer was asked to build a column that indicates whether "a participant consumed black jelly beans and white ones".

A Buggy Program

> eatBlackAndWhite =
    function(r):
      getValue(r, "black and white") == true
    end
> buildColumn(jellyAnon, "eat black and white", eatBlackAndWhite)

What is the Bug?

The logical and appears at a wrong place. The task is asking the programmer to write getValue(r, "black") and getValue(r, "white"), but the buggy program accesses the invalid column "black and white" instead.

A Corrected Program

> eatBlackAndWhite =
    function(r):
      getValue(r, "black") and getValue(r, "white")
    end
> buildColumn(jellyAnon, "eat black and white", eatBlackAndWhite)

pieCount

Context

jellyAnon

Task

The programmer was asked to visualize the proportion of participants getting acne.

A Buggy Program

> showAcneProportions =
    function(t):
      pieChart(count(t, "get acne"), "true", "get acne")
    end
> showAcneProportions(jellyAnon)

What is the Bug?

The program supplies a table produced by count to pieChart, which also expects a table and two of its column names. The table produced by count contains two column names, "value" and "count". Neither of the supplied colum names, "true" and "get acne", are column names of count(...).

A Corrected Program

> showAcneProportions =
    function(t):
      pieChart(count(t, "get acne"), "value", "count")
    end
> showAcneProportions(jellyAnon)

brownGetAcne

Context

jellyNamed

Task

The programmer was asked to compute how many participants consumed brown jelly beans and got acne, and how many did not.

A Buggy Program

> brownAndGetAcne =
    function(r):
      getValue(r, "brown") and getValue(r, "get acne")
    end
> brownAndGetAcneTable =
    buildColumn(jellyNamed, "part2", brownAndGetAcne)
> count(brownAndGetAcneTable, "brown and get acne")

What is the Bug?

The built column was named inconsistently. In buildColumn(...), the column was named "part2" but when counted, the column was accessed with "brown and get acne".

A Corrected Program

> brownAndGetAcne =
    function(r):
      getValue(r, "brown") and getValue(r, "get acne")
    end
> brownAndGetAcneTable =
    buildColumn(jellyNamed, "brown and get acne", brownAndGetAcne)
> count(brownAndGetAcneTable, "brown and get acne")

getOnlyRow

Context

students

Task

The programmer was asked to find Alice's favorite color.

A Buggy Program

> getValue(
    getRow(
      tfilter(students,
        function(r):
          getValue(r, "name") == "Alice"
        end),
      1),
    "favorite color")

What is the Bug?

There is only one row that matches the filtering criteria. So the only valid index is 0, not 1.

A Corrected Program

> getValue(
    getRow(
      tfilter(students,
        function(r):
          getValue(r, "name") == "Alice"
        end),
      0),
    "favorite color")

favoriteColor

Context

students

Task

The programmer was asked to define a function that finds all participants who like "green".

A Buggy Program

> participantsLikeGreen =
    function(t):
      tfilter(t,
        function(r):
          getValue(r, "favorite color")
        end)
    end

What is the Bug?

The programmer returns getValue(r, "favorite color") directly in the predicate but should return a boolean.

A Corrected Program

> participantsLikeGreen =
    function(t):
      tfilter(t,
        function(r):
          getValue(r, "favorite color") == "green"
        end)
    end

brownJellybeans

Context

jellyAnon

Task

The programmer was asked to count the number of participants that consumed jelly beans of a given color.

A Buggy Program

> countParticipants =
    function(t, color):
      nrows(tfilter(t, keep))
    end
> keep =
    function(r):
      getValue(r, "color")
    end
> countParticipants(jellyAnon, "brown")

What is the Bug?

"color" is not a valid column name. Instead of a string literal, the color should be a variable refering to the color defined in countParticipants.

A Corrected Program (1/2)

> countParticipants =
    function(t, color):
      nrows(tfilter(t, keep(color)))
    end
> keep =
    function(color):
      function(r):
        getValue(r, color)
      end
    end
> countParticipants(jellyAnon, "brown")

A Corrected Program (2/2)

> countParticipants =
    function(t, color):
      keep =
        function(r):
          getValue(r, color)
        end
      nrows(tfilter(t, keep))
    end
> countParticipants(jellyAnon, "brown")

employeeToDepartment

Context

• employees
• departments

Task

The programmer was given two tables, one maps employee names to department IDs, the other maps department IDs to department names. The task is to define a function, employeeToDepartment that consumes the two tables and looks up the a department name that an employee corresponds to.

A Buggy Program

> lastNameToDeptId =
    function(deptTab, name):
      matchName =
        function(r):
          getValue(r, "Last Name") == name
        end
      matchedTab = tfilter(deptTab, matchName)
      matchedRow = getRow(matchedTab, 0)
      getValue(matchedRow, "Department ID")
    end
> employeeToDepartment =
    function(name, emplTab, deptTab):
      buildColumn(emplTab, "Department Name",
        function(r):
          lastNameToDeptId(deptTab, getValue(r, "Last Name"))
        end)
    end

What is the Bug?

There are several problems in this program. First, employeeToDepartment is expected to return a department name, but it returns a table. Another problem is that the helper function is named lastNameToDeptId. The name suggests that this function maps the employee names to department IDs. But in employeeToDepartment, lastNameToDeptId is expected to produce department names. Finally, deptTab, the first parameter of lastNameToDeptId, has a name suggesting that it is bound to a department table. However, lastNameToDeptId uses deptTab as an employee table.

A Corrected Program

> deptIdToDeptName =
    function(deptTab, deptId):
      matchName =
        function(r):
          getValue(r, "Department ID") == deptId
        end
      matchedTab = tfilter(deptTab, matchName)
      matchedRow = getRow(matchedTab, 0)
      getValue(matchedRow, "Department Name")
    end
> employeeToDepartment =
    function(name, emplTab, deptTab):
      matchName =
        function(r):
          getValue(r, "Last Name") == name
        end
      matchedTab = tfilter(emplTab, matchName)
      matchedRow = getRow(matchedTab, 0)
      deptId = getValue(matchedRow, "Department ID")
      deptIdToDeptName(deptTab, deptId)
    end