PSeitz

For each segment: Open the `Column` via [FastFieldReaders](https://docs.rs/tantivy/latest/tantivy/fastfield/struct.FastFieldReaders.html). Access `ColumnValues` on `Column` and use [iter](https://docs.rs/tantivy-columnar/0.3.0/tantivy_columnar/column_values/trait.ColumnValues.html#method.iter) minimum and maximum values for numeric fast fields is already computed and accessible on ColumnValues

max aggregation on a string field?

You can have a look the docs here: https://docs.rs/tantivy/latest/tantivy/aggregation/index.html term_query would be `f1 = v1 AND f2=v2` and as collector the max aggregation

There's no multi-threading in this crate, but it would be generally possible.

Can you share more information about the query? Do you use the query parser or build the query yourself? Is your MUST condition a range query? There are two ways...

Do you use the fast field or the inverted index for your filter? The inverted index seems more suitable here

`fast` means a columnar index is created (fast is a misnomer), `indexed` creates the inverted index. The inverted index is already used for regular term queries.

Creating indices may cause problems, due to missing locking functionality on the file system. Otherwise it should be fine.

I think you would need to implement your own `Directory` without the locking calls

Yes, I had a look at that. `TwoPhaseIterator::approximation` returns a `DocIdSetIterator` where ``` The returned DocIdSetIterator is a superset of the matching documents, and each match needs to be confirmed...