Facettage.js < 05 | Examples list | 07 >

06 - Why (un-)serialization matters

WITHOUT SERIALIZATION

It generates an error (see console) because data is not serializable as a CSV.

WITH SERIALIZATION

It works. We should put that file in data/cache/ but it is already done.

Explanations

We load this data into facet without any serialize() or unserialize() methods:

[{
  name: 'George C. Scott',
  roles: [
    'Gen. "Buck" Turgidson'
  ]
},{
  name: 'Peter Sellers',
  roles: [
    'Lionel Mandrake',
    'President Merkin Muffley',
    'Dr. Strangelove'
  ]
}]

Remark that it cannot be properly represented as a CSV. The CSV conversion fails and it triggers an error.

Solution

We add a serialize() function to concatenate the list of roles into a string, allowing data to be transformed into a CSV. Of course a unserialize() function is required to do the reverse operation when data is retrieved from cache.

Note: using type: 'json' would solve the issue. It was for the sake of an example.

When using serialization?

When there is a type ambiguity, like a number that could be a string
When you have a Date or other objects
When you want to manage yourself how the data is stored, for instance to limit file size (Facettage does not zip data for you)