Populate XPath statements via variable

[mwiegand]

Good evening,

I was wondering if / how it’s possible to manage XPath declarations via variables. So to speak to dynamically define the amount of XPaths’ and for each a Columns name, XPath query, Type etc…

Thanks in advance
Mike

[janina]

Hallo @mw,

Yes, you can use Flow Variables to dynamically define XPath query, column names, etc. If you use a Chunk Loop Start you can also dynamically define the amount of XPaths. Here is a small example workflow to show you how it can be done:

XPath_Dynamic_Columns.knwf (18.4 KB)

I hope this answers your question.

Best,
Janina

[mwiegand]

Hi @janina,

apologize for my late reply and thanks a lot for your suggestion. The challenge I was referring to is the amount of XPaths as their corresponding flow variables become available only after the array of XPaths was declared.

In other words, how can an variating amount of XPath declarations be configured via flow variables? The only approach I see is:

1. Manually define a fixed and large amount of xPaths (i.e. 100)
2. Dynamically define and XPath declarations per your suggestion
3. Accept a messy XPath node configuration

However, this still requires a lot of manual configuration steps basically neglecting the automation approach.

@ipazin Wouldn’t it be a great feature to pass an array pf XPaths like so to the XPath node to automate the node setup? Maybe this is “hidden” behind the flow variable "numberOfQueries), though?

Column name	Column name is XPath	XPath query	Return Type	Return missing on empty	Multiple tag options
Fixed title	false	//book/title	String cell	false	Single Cell
//title	true	//book	Node cell	false	Single Cell

The array index could dynamically be determined by the row line or manually set if necessary. In case a JSON is required the table should be able to get converted automatically too. What do you think?

The scenario / benefits would be:

1. Auto Separate / Sort XML nodes into column based on XML information
2. Configure all XPath declarations for entire workflow and switch based on conditions.
   I.e. when XML documents with different structures must be processed
3. Automate XPath setup.

Here is an example workflow based on that from @janina which explores the automation further.

Xpath dynamically define queries – mw

Please note that, when trying to configure the “Return type” and “Multi tag options”, regardless of using the String values or indexes, I got the following errors:

WARN  XPath                3:539      Errors loading flow variables into node : Coding issue: No enum constant org.knime.xml.node.xpath2.XPathNodeSettings.XPathMultiColOption.Single Cell
WARN  XPath                3:539      Errors loading flow variables into node : Coding issue: No enum constant org.knime.xml.node.xpath2.XPathNodeSettings.XPathMultiColOption.0

This resulted in failure when trying to retrieve the column names dynamically in “Row6”:

Best
Mike