Struct QueryApi 

pub struct QueryApi<S: StorageClient + 'static> { /* private fields */ }

Universal Query API – Core Layer

Implementations

impl<S: StorageClient + Clone + 'static> QueryApi<S>

pub fn new(storage: Arc<RwLock<S>>, stats_manager: Arc<StatsManager>) -> Self

Create a new QueryApi instance with external StatsManager

pub fn execute_with_execution( &mut self, query: &str, ctx: QueryRequest, execution: &TransactionExecution, ) -> CoreResult<QueryResult>

Execute a query with an explicit transaction execution binding.

The execution parameter carries the full transaction identity (ID, timestamps, mode, owner) from TransactionManager. This is the preferred entry point for all transactional DML.

pub fn execute( &mut self, query: &str, ctx: QueryRequest, ) -> CoreResult<QueryResult>

Execute a query with the given query request

Parameters

query: The query statement

• ctx: query request

Return

Structured Search Results

pub fn execute_with_operation_storage( &mut self, query: &str, ctx: QueryRequest, storage: S, ) -> CoreResult<QueryResult>

pub fn execute_stream_with_execution( &mut self, query: &str, ctx: QueryRequest, execution: &TransactionExecution, ) -> CoreResult<StreamingQueryResult>

Execute a query with an explicit transaction execution binding (streaming).

pub fn execute_stream( &mut self, query: &str, ctx: QueryRequest, ) -> CoreResult<StreamingQueryResult>

Execute a query and return a StreamingQueryResult for chunk-at-a-time consumption.

Unlike [execute] which materialises the full result set, this method returns a thread-safe streaming handle that lets the caller pull chunks one at a time. Useful for SSE / gRPC streaming endpoints.

pub fn execute_stream_with_operation_storage( &mut self, query: &str, ctx: QueryRequest, storage: S, ) -> CoreResult<StreamingQueryResult>

impl<S> QueryApi<S>

where S: StorageClient + Clone + AutoCommitBatchOps + 'static,

pub fn execute_batch( &mut self, queries: &[String], ctx: QueryRequest, ) -> Vec<Result<QueryResult, CoreError>>

pub fn dml_plan_memo_hits(&self) -> u64

pub fn dml_template_ast_parse_count(&self) -> u64

pub fn dml_template_ast_hit_count(&self) -> u64

impl<S> QueryApi<S>

where S: StorageClient + Clone + AutoCommitBatchOps + AutoCommitGroupOps + 'static,

pub fn execute_batch_grouped( &mut self, queries: &[String], ctx: QueryRequest, group_size: usize, ) -> Vec<Result<QueryResult, CoreError>>

Execute a batch of auto-commit DML statements in group-commit windows. group_size is clamped to >= 1; each consecutive group of statements shares one write timestamp, one WAL fsync, and one commit point. Statement results are returned per statement, in order, exactly like execute_batch; failed statements roll back only their own writes. DDL / transaction statements / USE must not be mixed in (same contract as execute_batch).

impl<S: StorageClient + Clone + 'static> QueryApi<S>

pub fn with_schema_manager(self, schema_manager: Arc<SchemaManager>) -> Self

Attach a schema manager to the pipeline.

pub fn with_sync_manager(self, sync_manager: Arc<SyncManager>) -> Self

Attach a sync manager to the pipeline.

pub fn with_optimizer_engine( storage: Arc<RwLock<S>>, stats_manager: Arc<StatsManager>, optimizer_engine: Arc<OptimizerEngine>, schema_manager: Option<Arc<SchemaManager>>, ) -> Self

Create a new QueryApi using an externally configured optimizer engine, so server-level settings (e.g. the [parallel] partitioning section) reach the query pipeline. The engine instance is shared by value.

pub fn with_cost_config(self, config: CostModelConfig) -> Self

Apply an explicit cost model configuration, replacing the default engine built by the constructors.

Assembly-time only: call before the first query. The replacement engine is freshly built, exactly like the constructor default; previously installed per-space overrides are carried over.

pub fn with_cost_profile( self, profile: StorageCostProfile, runtime: &RuntimeConfig, ) -> Self

Apply a storage cost profile with runtime knowledge.

Memory-mode deployments short-circuit to the in-memory preset without touching the filesystem; embedded in-memory users must call this (or Self::with_cost_config), otherwise the constructor default prices queries with mechanical-drive costs.

pub fn with_space_cost_profiles( self, profiles: &HashMap<String, StorageCostProfile>, runtime: &RuntimeConfig, ) -> Self

Apply per-space storage cost profile overrides.

Spaces absent from the map keep the global profile; memory-mode deployments short-circuit every space to the in-memory preset. Assembly-time only: call before the first query.

pub fn with_default_cost_profile_label(self, label: String) -> Self

Override the global cost profile display label (assembly-time only).

Used by assemblies that build their own engine (e.g. the network server with extra engine settings) so EXPLAIN still names the effective profile.

pub fn cost_config(&self) -> CostModelConfig

Effective cost model configuration currently in force.

pub fn with_statistics_settings( self, sample_limit: usize, min_epoch_delta: u64, ) -> Self

Apply the server-configured statistics collection settings to the pipeline (sampling window and automatic-refresh epoch delta).

The pipeline keeps fallback defaults when this is never called, so embedded users without server config are unaffected.

pub fn statistics_sample_limit(&self) -> usize

Effective statistics sampling window currently in force.

pub fn with_password_history_depth(self, depth: usize) -> Self

Set the retained password-history depth applied by user-management operations (assembly-time only).

pub fn statistics_min_epoch_delta(&self) -> u64

Effective statistics refresh epoch delta currently in force.

pub fn install_shared_scheduler( &mut self, shared_scheduler: Arc<SharedScheduler>, query_registry: Arc<QueryRegistry>, )

Install a shared scheduler and query registry into an existing pipeline (used when the pipeline was built by another constructor).

pub fn shared_scheduler(&self) -> Option<Arc<SharedScheduler>>

Access the shared scheduler instance held by the pipeline, if any.

pub fn query_registry(&self) -> Option<Arc<QueryRegistry>>

Access the query registry instance held by the pipeline, if any.

pub fn install_query_manager(&mut self, query_manager: Arc<QueryManager>)

Install the global QueryManager so per-query runtimes forward KILL QUERY, finish tracking, and progress events into one registry. Mirror of [install_shared_scheduler]: called once at server startup.

pub fn set_extensions(&mut self, extensions: Arc<ExtensionRegistry>)

Install a shared pipeline-extension registry (unstable experimental API) into an existing pipeline.

pub fn extensions(&self) -> Arc<ExtensionRegistry>

Pipeline-extension registry backing this API.

pub fn collect_statistics(&self, space: &str, force: bool) -> CoreResult<()>

Collect (or serve cached) optimizer statistics for a space.

force bypasses the schema-version gate so an explicit ANALYZE always refreshes the statistics. Failures are returned as Err, never panic.

pub fn plan_cache_hit_rate(&self) -> f64

Query-plan-cache hit rate across all executed statements.

pub fn plan_cache_len(&self) -> usize

Number of entries currently held in the query plan cache.