Skip to content

Columnar Analytics

Uni executes filters, projections, and aggregations using columnar, vectorized execution (Apache Arrow + DataFusion). This gives fast analytical queries without a separate warehouse.

What It Provides

  • Vectorized scans and predicate pushdown.
  • Aggregations like COUNT, SUM, AVG, COLLECT.
  • Columnar storage via LanceDB.
  • Windowed analytics via OVER (...) clauses (see Window Functions).

Example

use uni_db::Uni;

# async fn demo() -> Result<(), uni_db::UniError> {
let db = Uni::open("./my_db").build().await?;

let results = db.query(
    "MATCH (p:Purchase) RETURN p.category, SUM(p.amount) AS revenue"
).await?;

println!("{:?}", results);
# Ok(())
# }

import uni_db

db = uni_db.Database("./my_db")
rows = db.query(
    "MATCH (p:Purchase) RETURN p.category, SUM(p.amount) AS revenue"
)
print(rows)

Use Cases

  • Product or sales analytics in an application.
  • Aggregations over large property sets.
  • Fast filters and grouping without a separate analytics system.

When To Use

Choose columnar analytics when you need fast aggregations and scans over large datasets, but still want graph and vector queries in the same system.