The search tress would be used to return partially matched lists from the database, i.e. for autocomplete and similar functionality. And it should have a lesser memory footprint than standard prefix trees.

What exactly would you use the ternary search trees for?
Full-text searches? And if yes, do you have any experience with how the memory consumption of a ternary tree would be compared to regular prefix trees? Would be good to know!

Namely looking toward this item on the roadmap: “Graphs:
Advanced Algorithms”.

Honestly though, what takes an even higher priority for my projects is ternary search trees. Thank you for asking, and thank you for hosting this superb software.

Hi Maxwell,

are you waiting for something in particular? I am asking because the next release focuses on replication, but we are more than happy to hear your requirements on the graph api so that we can consider it either for this release or for the roadmap.