Solr Numerical Trie vs traditional trie (prefix tree)

Question

I'm reading Taming Text by Ingersoll, Morton & Farris and I don't get how solr's implementation of numerical trie would help search for text? I'm a bit confused with the explanation for solr.TrieField fieldType for solr.

What I am confused is: Traditional prefix tree stores alphabet (if it's English), solr's is storing numerical, I don't get how someone is able to search for text in a numerical tree.

Does the tree convert characters into number and then range it?

It's apparently a range field. From what I've gathered, say we type a prefix Tamp, then from there we can stem to different possibility and a range query works with prefix + [range of character in numeric]?

Thank you for you time

Answer 1

Solr support searching numeric fields for numeric ranges. It is useful for things like price facets.

Numeric Trie generates the range boundaries at different levels of precision, so when you are looking for the entries that fit within the range, you could reject whole groups of them at once (e.g. anything below 2^5 is definitely not within range).