Export As YAML

A demonstration on how to use export templating to produce YAML.

Introduction

For this tutorial, we are going to be working with a small set of data that looks like this:

Name	Height (m)	Occupation
Andy	1.73	Student
Bernd	1.79	Zoologist
Carol	1.65	Carpenter

By default, OpenRefine exports JSON. JSON is a very effective way to transport data, but it can sometimes be difficult to read. Consider the difference between these two presentations of the same data:

JSON

{
  "rows" : [
    {
      "Name" : "Andy",
      "Height (m)" : 1.73,
      "Occupation" : "Student"
    },
    {
      "Name" : "Bernd",
      "Height (m)" : 1.79,
      "Occupation" : "Zoologist"
    },
    {
      "Name" : "Carol",
      "Height (m)" : 1.65,
      "Occupation" : "Carpenter"
    }
  ]
}

YAML

rows:

- Name: Andy
  Height (m): 1.73
  Occupation: Student

- Name: Bernd
  Height (m): 1.79
  Occupation: Zoologist

- Name: Carol
  Height (m): 1.65
  Occupation: Carpenter

You can see from these two examples that they are both fairly readable. However, YAML has no commas or curly braces that interfere with readability.

Getting Started

You can follow along by pointing OpenRefine at the example spreadsheet:

• Go to http://localhost:3333/ in your web browser
• Begin a new project with https://spreadsheets.google.com/pub?key=0Aq_3OYelM4ZUdEpYYmNKY0t1V0VjYlR5UjFYejVHbWc&hl=en_GB&single=true&gid=0&range=A1%3AC4&output=csv as the URL.
• Click on Export > Templating...

Details

YAML is a very simple format. Its entire purpose is to be human readable. Also, it is focused on data, rather than documents. This makes it a perfect fit for sharing data from OpenRefine.

Changes to make

Prefix Field

Basic We only need a single word in the prefix field:

rows:

Bonus points

If we want to be nice to YAML parsers, we can add some optional dashes. These help to tell the computer that's reading it that it has entered a YAML document. The pound sign or octothorpe (#) is the comment character.

--- #people

rows:

Row Field

Generating rows is much simpler than the current JSON. The main thing to watch out for is that there is a blank line between rows.

Try changing:

    {
      "Person" : {{jsonize(cells["Person"].value)}},
      "Height (m)" : {{jsonize(cells["Height (m)"].value)}},
      "Occupation" : {{jsonize(cells["Occupation"].value)}}
    }

To:

-  Name: {{cells["Person"].value}}
   Height (m): {{cells["Height (m)"].value}}
   Occupation: {{cells["Occupation"].value}}

Some points here:

• Keep the curly braces in place. That tells OpenRefine that you need to do some processing there.
• Remove the trailing commas from each part
• Make sure the cell headings match exactly. If you see "null" in the preview window instead of the correct value, you probably have the column name wrong.

Row Separator

Change the row separator from the default comma (,) to a blank line. Confirm that there's a blank line between each record in the preview window.

Suffix Field

Adding three dots is good practice. It tells the parser that we've finished.

...