Currently we supported column filter for properties but not node and sub-node level. For node, only filter by node identifier which does not make sense (MGNLUI-5903).
This cause some difficulties when customers using filter with nodes.

The improvement concentrate on supporting easy of use and flexibility of configuration in that:

• Allow customer to configure filter property name for nodes. In case the filter works on properties, the filter property name should be ignore or directly use the filtering property as the filter property name (which allow consistency in implementation).
• Development notes:
  • From this pull request we can see that we should take the configured filter property name to get the value to filter instead of use the node identifier.
  • Also in case the filter property name does not exist, do not stop the processing flow. Use a lowest priority value for it instead with a warning or info in the log.
  • We may not support sub-nodes filtering to reduce the usage complexity and implementation efforts. With a note that we are trying to simplify the product to help customers easier to adopt and use it rather than supporting a lot of complexity functions. Customers who want to have more specific functions should extends our implementation or request our Professional services for doing that.