Example for building an external index for parquet files

[alamb]

Is your feature request related to a problem or challenge?

It is common in databases and other analytic system to have additional external "indexes" (perhaps stored in the "metadata catalog", perhaps stored alongside the data files, perhaps embedded in the files, perhaps elsewhere)

These indexes are used to speed up queries by "pruning" the files -- specifically evaluating a predicate on the index and then only reading files / portions of files that would pass the filters in the query

Implementing such a index requires several three steps:

1. Creating / maintaining the actual index
2. A way to evaluate the index against the query predicate
3. A way to use the index to filter the files that need to be read

DataFusion has code already to do 2 (PruningPredicate) and 3 (ParquetExec) but I am not sure how obvious it is to put togeher

DataFusion actually also has basic support for 1 (e.g. the ListingTableProvider reads statistics and prunes based on their statistics)

We have some version of this in InfluxDB today, and we are in the process of extending it. However, I think the usecase is much more general and could be useful for other systems as well (e.g. full text indexes across parquet, for example).

I also think creating an example will help motivate the work from @NGA-TRAN in #10453 and myself in #9929

Describe the solution you'd like

No response

Describe alternatives you've considered

No response

Additional context

No response

[alamb]

Update here is I have a basic example #10549 ready for review / merge

[adriangb]

Sorry for jumping in here, maybe this isn't the best issue but it's hard to keep up with all of the amazing work you're doing @alamb!

I wanted to pitch a use case I've been thinking about of storing a secondary index on a searchable async location. Think a relational database with ACID guarantees. In particular the key would be that hooks to do selections / pruning be async and that they pass in filters: I'd push down the filters into filters in the metadata store and run an actual query there that returns the files / row groups to scan. This is in contrast to #10549 for example where the index is in memory and fully materialized. I realize that TableProvider.scan already serves this purpose, but it'd be nice to integrate into these new APIs instead of having to implement more things oneself because you're hooking in at a higher (lower?) level.