Struct StorageEngine 

pub struct StorageEngine { /* private fields */ }

Storage engine for Arrow/Parquet data

Implementations

impl StorageEngine

pub const fn new(batches: Vec<RecordBatch>) -> Self

Create a new storage engine from existing batches

Useful for testing and benchmarking

pub fn load_parquet<P: AsRef<Path>>(path: P) -> Result<Self>

Load table from Parquet file

Errors

Returns error if file cannot be read or parsed

pub fn batches(&self) -> &[RecordBatch]

Get all record batches

pub fn morsels(&self) -> MorselIterator<'_>

Create iterator over morsels (128MB chunks)

pub fn append_batch(&mut self, batch: RecordBatch) -> Result<()>

Append batches to storage (OLAP-optimized)

WARNING: This is the ONLY supported write operation. Trueno-DB does NOT support incremental row updates (OLTP).

Design Rationale

Columnar storage optimizes for bulk reads, not random writes:

• Single-row update cost: O(N) (rewrite entire column)
• Batch append cost: O(1) (append to new partition)

Example

// Good: Bulk append (OLAP-compatible)
let schema = Arc::new(Schema::new(vec![
    Field::new("id", DataType::Int32, false),
]));
let batch = RecordBatch::try_new(
    schema,
    vec![Arc::new(Int32Array::from(vec![1, 2, 3]))],
)?;

let mut storage = StorageEngine::new(vec![]);
storage.append_batch(batch)?;

Errors

Returns error if batch schema doesn’t match existing batches

pub fn update_row(&mut self, _row_id: usize, _values: RecordBatch) -> Result<()>

👎Deprecated since 0.1.0: Trueno-DB is OLAP-only. Use append_batch() for bulk data loads.

DEPRECATED: Single-row update not supported

Trueno-DB is OLAP-only (columnar storage). Use append_batch instead.

Why This Fails

Column stores are optimized for bulk reads, not random writes:

• SQLite (row-store): O(1) update with B-tree index
• Trueno-DB (column-store): O(N) update (rewrite entire column)

Migration Guide

// Bad: Incremental update (OLTP pattern)
// storage.update_row(row_id, new_values)?;  // NOT SUPPORTED

// Good: Batch re-analysis (OLAP pattern)
let schema = Arc::new(Schema::new(vec![
    Field::new("id", DataType::Int32, false),
]));
let new_batch = RecordBatch::try_new(
    schema,
    vec![Arc::new(Int32Array::from(vec![1, 2, 3]))],
)?;
let mut storage = StorageEngine::new(vec![]);
storage.append_batch(new_batch)?;

Errors

Always returns error (not implemented)