kandra-runtime/io.kandra.runtime/QueryExecutor

QueryExecutor

class QueryExecutor(session: CqlSession, schema: TableSchema, statementBuilder: StatementBuilder, codec: KandraCodec = KandraCodec.default, debugConfig: DebugConfig = DebugConfig(), maxUnpagedResultRows: Int)

Translates QueryContext predicates into CQL SELECT statements and decodes results.

Lookup predicates trigger a two-step query:

Query the lookup table for the primary table's full key (partition + clustering columns).
Query the primary table by that full key.

@SecondaryIndex predicates query the primary table directly (no two-step needed). A WARN is logged each time a secondary index query executes.

IN on a single-column partition key is supported — a DEBUG message is logged since it causes scatter-gather across partitions (legitimate but worth tracking).

Constructors

QueryExecutor

constructor(session: CqlSession, schema: TableSchema, statementBuilder: StatementBuilder, codec: KandraCodec = KandraCodec.default, debugConfig: DebugConfig = DebugConfig(), maxUnpagedResultRows: Int)

Functions

exists

fun exists(consistency: KandraConsistency? = null, block: QueryContext.() -> Unit): Boolean

existsSuspend

suspend fun existsSuspend(consistency: KandraConsistency? = null, block: QueryContext.() -> Unit): Boolean

find

fun <T : Any> find(entityClass: KClass<T>, consistency: KandraConsistency? = null, block: QueryContext.() -> Unit): T?

Like findAll but only ever needs the first matching row — passes limitOne = true into resolveRows (GH #106 / ISS-093) so the direct-CQL branch appends LIMIT 1 instead of fetching up to maxUnpagedResultRows rows over the wire just to take the first. The IN and lookup-table branches are unaffected (limitOne isn't meaningful there — a lookup match is already exactly one row, and an IN scatter-gather ignores it, same as before).

findActive

fun <T : Any> findActive(entityClass: KClass<T>, allowFullScan: Boolean = false): List<T>

findActiveSuspend

suspend fun <T : Any> findActiveSuspend(entityClass: KClass<T>, allowFullScan: Boolean = false): List<T>

findAll

fun <T : Any> findAll(entityClass: KClass<T>, consistency: KandraConsistency? = null, block: QueryContext.() -> Unit): List<T>

findAllSuspend

suspend fun <T : Any> findAllSuspend(entityClass: KClass<T>, consistency: KandraConsistency? = null, block: QueryContext.() -> Unit): List<T>

findById

fun <T : Any> findById(entityClass: KClass<T>, vararg idValues: Any, consistency: KandraConsistency? = null): T?

findByIdSuspend

suspend fun <T : Any> findByIdSuspend(entityClass: KClass<T>, vararg idValues: Any, consistency: KandraConsistency? = null): T?

findPage

fun <T : Any> findPage(entityClass: KClass<T>, pageSize: Int, pageToken: String?, consistency: KandraConsistency? = null, block: QueryContext.() -> Unit): KandraPage<T>

findPageSuspend

suspend fun <T : Any> findPageSuspend(entityClass: KClass<T>, pageSize: Int, pageToken: String?, consistency: KandraConsistency? = null, block: QueryContext.() -> Unit): KandraPage<T>

findSuspend

suspend fun <T : Any> findSuspend(entityClass: KClass<T>, consistency: KandraConsistency? = null, block: QueryContext.() -> Unit): T?

Suspend counterpart of find — see its doc (GH #106 / ISS-093).

raw

fun raw(cql: String, vararg params: Any?): List<Row>

rawQuery

fun rawQuery(query: KandraRawQuery): List<Row>

rawQuerySuspend

suspend fun rawQuerySuspend(query: KandraRawQuery): List<Row>

rawSuspend

suspend fun rawSuspend(cql: String, vararg params: Any?): List<Row>