Native XML databases - Making Sense of NoSQL - page 98

Like the JSON (JavaScript Object Notation) data format (a text-based standard for
human-readable data interchange), XML stores its information in a hierarchical tree
structure. Unlike JSON , XML can store documents with mixed content and
namespaces. Mixed-content systems allow you to mix sequences of text and data in
any order. Elements in XML can contain text that has other trees of data interspersed
throughout the text. For example, you can add HTML links that contain bold, italic,
links, or images anywhere inside a paragraph of text. HTML files are a perfect example
of the type of mixed content that can't be stored by or queried in a document data-
base that only supports the JSON format.
Figure 5.1 shows the relationship between the features of comma-separated value
( CSV ) flat files used in many SQL systems, JSON , and XML documents.
If you use spreadsheets or load data into RDBMS tables, you know that CSV files are
an ideal mechanism for storing data that's loaded into a single table. The CSV struc-
ture of commas separating fields and newline characters separating rows is frequently
used to transfer data between spreadsheets and RDBMS tables.
JSON files are ideal for sending serialized objects to and from a web browser. JSON
allows objects to contain other objects, and works well in hierarchical structures that
don't need to store mixed content or use multiple namespaces. Due to its familiarity
in the JavaScript world, JSON is the de facto standard for storing hierarchical docu-
ments within a document store. But it was never designed as a general-purpose con-
tainer for mixed-content markup languages such as HTML .
As mentioned, XML files are the best choice when your document includes mixed
content. XML also supports an often-controversial feature: namespaces. Namespaces
allow you to mix data elements from different domains within the same document, yet
retain the source meaning of each element. Documents that support multiple
namespaces allow applications to add new elements in new namespaces without dis-
rupting existing data queries.

Extensible
markup language
(XML)

XML

Mixed content

JSON

Namespaces

Mixed

JavaScript
object notation
(JSON)

Hierarchical
documents

CSV

Comma-separated
value (CSV)

Flat files

Figure 5.1 The expressiveness of three document formats. Comma-separated value
(CSV) files are designed to only store flat files that don't contain hierarchy. JavaScript
Object Notation (JSON) files can store flat files as well as hierarchical documents.
Extensible Markup Language (XML) files can store flat files, hierarchical documents,
and documents that contain mixed content and namespaces.