Working with XML Documents

[oswaldobapvicjr]

I have an explicit requirement to filter a previously loaded XML document using an XPath expression in a custom function. For example:

xpath(document, "/products/product[1]/price")

However, to do it in EvalEx, I would need the EvaluationValue class to accept the DataType.DOCUMENT, stored as org.w3c.dom.Document from the Java API (no additional dependency required). Could you consider having it in EvalEx?

A choice would be accepting a more generic DataType type (e.g.: DataType.UNDEFINED). The point is to take the Document as a variable inside the context so that the xpath function can filter it. This second choice feels more "open" to further requirements. Since EvaluationValue.DataType is an enum, it doesn't allow adding custom data types on top of it without changing the class.

Note 1: We could also accept a org.w3c.dom.Node, since a Document is a type of (root) Node.

Note 2: Inspired by the usability of MSExcel function: FILTERXML(xml, xpath)

[uklimaschewski]

I think that a DataType.BINARY would be an aceptable generic new data type for such cases.
The new converter could be last in the converter chain in DefaultEvaluationValueConverter, allowing to convert any object that wasn't converted before (just taking the Object and setting it as value).

[oswaldobapvicjr]

Good. I think I can implement it with a pull request.

[oswaldobapvicjr]

I just created the Pull Request #482 with this implementation. ;)