You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
At the moment if you supply a labelFilter any non-listed labels are assumed to be black listed. I think having an option to assume that all non listed labels are white listed and only specify black listed labels / end labels would be a benefit to the function (or its config cousin).
The text was updated successfully, but these errors were encountered:
Comment by InverseFalcon Thursday Mar 16, 2017 at 01:17 GMT
I'm guessing that this was submitted with the assumption that the labelFilter supports multiple operations at once (+, -, /) instead of only one at a time.
Currently the label filter only allows one operation, the one specified on the very first label of the filter.
This means for a blacklist filter, anything not in the blacklist is allowed, and for the end node label filter, anything not in the end node filter is allowed as a component of a path leading up to an end node, but will never be the end node of a returned path.
#317 is a proposed improvement to allow multiple operations in the filter at once, and addresses the idea that if using an end-node label filter or a blacklist label filter (either alone or in combination with each other), that an empty whitelist should be interpreted as allowing all instead of disallowing all.
Otherwise, if there is a non-empty whitelist label filter, it should probably still exclude any nodes not in the whitelist.
There is still some question regarding the order of evaluation for all three types of filters. If we decide, for example, that the end-node label filter should evaluate before the whitelist, then a node with a label in the end-node label filter should be included even if that label is not in the whitelist, as the whitelist won't have a chance to evaluate it.
There is still ambiguity regarding multi-labeled nodes where its labels fall within multiple label filters, that's up for discussion in #317 .
Comment by InverseFalcon Tuesday Mar 21, 2017 at 01:23 GMT
To summarize, I don't think this is a real issue in current implementation, and the concern is being addressed in the work done on allowing multiple label filter operations instead of just one.
Issue by dan-kez
Thursday Oct 20, 2016 at 16:24 GMT
Originally opened as neo4j-contrib/neo4j-apoc-procedures#181
At the moment if you supply a labelFilter any non-listed labels are assumed to be black listed. I think having an option to assume that all non listed labels are white listed and only specify black listed labels / end labels would be a benefit to the function (or its config cousin).
The text was updated successfully, but these errors were encountered: