Check for Coding Errors¶

Datavyu scripts have the ability to check for coding errors to make up for potential human error. Inputting invalid codes is a common coding mistake. For example, a coder might accidentally input an “h”, when only “j” or “k” are acceptable values. Using scripts, coders can double-check their work for errors and fix them early on.

Datavyu API provides the checkValidCodes() method to check for coder errors. checkValidCodes() requires at least three arguments: the name of the column it will verify, the location that it will output the results to, and at least one key-value pair. Key-value pairs consist of a “key”, the name of a Datavyu code, and a “value”, which is an array of valid values for that code. checkValidCodes() then checks each code (key) against its list of valid values (the values).

The following examples check the look column against its codes. You can download the sample data used in this tutorial from here.

Check Code Validity and Output to the Console¶

Basic Format¶

Example

require 'Datavyu_API.rb'
begin
   # Check for errors. Notice the square brackets. These denote
   # lists. The basic format is: "columnName", "dumpFile",
   # "codename", ["validcode1", "validcode2", etc],
   # "code2", [ "validcode1", "validcode2", etc], ...

   checkValidCodes("look", "", "direction", ["l", "r"], "target", ["a", "b"])
end

Breaking down the function call makes it easier to follow what is happening. Recall that checkValidCodes() takes (at least) three arguments: the name of the Datavyu variable it will be checking, the location that it should direct the output to, and at least one key-value pair.

In the example, we have:

checkValidCodes("look", "", "direction", ["l", "r"], "target", ["a", "b"])

"look" is the name of the variable to check.
"" is the location that we want to export the output to. Using "" indicates that we do not want to write the results to a file, and that it should instead display in the Datavyu Scripting Console.
"direction", ["l", "r"] is the first key-value pair, which specifies that the “direction” Datavyu code should only have the values “l” or “r”.
"target" , ["a", "b"] is the second key-value pair, which specifies that the “target” Datavyu code should only have the values “a” or “b”.

Advanced Format¶

You can perform the same verification by first assigning the valid codes to objects. This makes it is easier for human readers to parse the script, and makes it easy to modify or update in the future.

Set up the script:
```
require 'Datavyu_API.rb'
begin
```

Assign each list of valid codes to a variable:

require 'Datavyu_API.rb'
begin
   # Store each of the valid code arrays into a object first
   # so that it is easier to read
   directionCodes = ["l", "r"]
   targetCodes = ["a", "b"]

Check for coding errors using checkValidCodes(), replacing the lists with your newly-created objects, and end the script:

require 'Datavyu_API.rb'
begin
   directionCodes = ["l", "r"]
   targetCodes = ["h", "t"]

   # Check for typos, replacing the code arrays with your new variables:
   checkValidCodes("look", "", "direction", directionCodes, "target", targetCodes)
end

Check Code Validity and Output to a File¶

checkValidCodes() can write the results of its verification to a “dump file”. Datavyu will create an output file if you specify a path in the dumpFile parameter.

Create an object that holds the path that it should output to using Ruby’s File.expand_path method, which converts a relative path, like ~/Desktop/file.txt to an absolute path name, which contains the root directory, and all sub-directories, like /Users/alice/Desktop/file.txt

The following commands create a variable called output that holds the absolute path to a file on the Desktop called output.txt:
```
require 'Datavyu_API.rb'
begin

   output = File.expand_path("~/Desktop/output.txt")
```

Call checkValidCodes() on the “step” column, passing output as the argument for the dumpFile parameter:

require 'Datavyu_API.rb'
begin
   directionCodes = ["l", "r"]
   targetCodes = ["a", "b"]

   output = File.expand_path("~/Desktop/output.txt")

   # Check for errors, specifying the output variable as the dumpFile parameter
   checkValidCodes("look", output, "direction", directionCodes, "target", targetCodes)
end

When the script ends, the output.txt file will be created on the Desktop, containing the results of the code checking. For the sample data, it should find one error, and the output should resemble:

Code ERROR: Var: look   Ordinal: 2      Arg: direction     Val: rj

Video Example of Checking for Errors¶

This video displays one way to check for errors (typos, impossible values, etc.) within a spreadsheet.

← Add Cells to a Column

Check Inter-rater Reliability to Improve Data Accuracy →