Data content, processing

By default, the data content of an XML or SGML document is streamed through to the current output scope by the parser. You can intercept and process data content in one of four ways:

Using data-content rules

If you add a data-content rule to your program, it will be fired whenever a continuous piece of text data occurs in your input data. You can then process that text by scanning "%c":

  data-content
     repeat scan "%c"
        ...
     again

You can restrict a data-content rule to a particular element by adding a condition to the rule:

  data-content when element is "product-name"
     repeat scan "%c"
        ...
     again

A data-content rule processes a contiguous sequence of text characters. A contiguous sequence of text characters is bounded by:

  • the start of an element
  • the end of an element
  • a processing instruction
  • an external CDATA, SDATA, NDATA, or SUBDOC entity reference

Using translate rules

If you put translate rules into your program they will scan data-content (and attribute content) automatically, without the need for you to explicitly initiate scanning. In effect, translate rules work like find rules, except that they are initiated by do xml-parse or do sgml-parse instead of submit.

  translate "$" digit+ => dollars
            ("." digit{2} => cents)?
  
     output dollars
     output "," || cents when cents is specified
     output "$"

When processing SGML, you can also use translate rules to capture and process entities.

Scanning "%c" in an element rule

You can also process data-content by scanning %c in an element rule or another markup rule. However, you should be aware that such a scanning process will scan the result of all the parsing operations that take place on the content of an element, including the processing of any element, translate, or data-content rules, not on the raw data content of the element.

You should scan "%c" only if you know that the current element contains only data content or you want to scan the result of parsing the current element. Bear in mind that even if the element has only data content, any applicable translate rules and data content rules will fire before the scanning operation takes place, and the scanning source will be the output of those rules acting on the data content, not the raw data content of the element.

Processing #content in a markup rule

While %c enables another pass through already-processed content, #content provides access to unprocessed content. The downside of using #content for processing data content is that, besides the plain textual content, it may contain unprocessed markup events: the type of #content is markup source, not string source.

Before you can scan or submit #content, you first need to convert it to a string source using the operator take, as in the following data-content rule:

  data-content
     repeat scan #content take any*
        ...
     again

You should be aware, however, that the take operator will throw any markup event it encounters in #content, such as an SDATA entity reference. You should always wrap applications of take and drop operators to a markup source in a scope that catches all markup event throws. The following rule will process all data content inside an element, while ignoring all element tags and other markup:

  element #implied
     repeat
        submit #content take any*
        exit
  
      catch #markup-point
      catch #markup-start
      catch #markup-end
     again