Package org.apache.calcite.config

Class CalciteSystemProperty<T>

java.lang.Object
org.apache.calcite.config.CalciteSystemProperty<T>
Type Parameters:
T - the type of the property value
public final class CalciteSystemProperty<T> extends Object
A Calcite specific system property that is used to configure various aspects of the framework.

Calcite system properties must always be in the "calcite" root namespace.

  • Field Details

    • DEBUG

      public static final CalciteSystemProperty<Boolean> DEBUG
      Whether to run Calcite in debug mode.

      When debug mode is activated significantly more information is gathered and printed to STDOUT. It is most commonly used to print and identify problems in generated java code. Debug mode is also used to perform more verifications at runtime, which are not performed during normal execution.

    • COMMUTE

      public static final CalciteSystemProperty<Boolean> COMMUTE
      Whether to exploit join commutative property.

    • ENABLE_COLLATION_TRAIT

      public static final CalciteSystemProperty<Boolean> ENABLE_COLLATION_TRAIT
      Whether to enable the collation trait in the default planner configuration.

      Some extra optimizations are possible if enabled, but queries should work either way. At some point this will become a preference, or we will run multiple phases: first disabled, then enabled.

    • ENABLE_ENUMERABLE

      public static final CalciteSystemProperty<Boolean> ENABLE_ENUMERABLE
      Whether the enumerable convention is enabled in the default planner configuration.

    • ENUMERABLE_ENABLE_TABLESCAN_ARRAY

      public static final CalciteSystemProperty<Boolean> ENUMERABLE_ENABLE_TABLESCAN_ARRAY
      Whether the EnumerableTableScan should support ARRAY fields.

    • ENUMERABLE_ENABLE_TABLESCAN_MAP

      public static final CalciteSystemProperty<Boolean> ENUMERABLE_ENABLE_TABLESCAN_MAP
      Whether the EnumerableTableScan should support MAP fields.

    • ENUMERABLE_ENABLE_TABLESCAN_MULTISET

      public static final CalciteSystemProperty<Boolean> ENUMERABLE_ENABLE_TABLESCAN_MULTISET
      Whether the EnumerableTableScan should support MULTISET fields.

    • ENABLE_STREAM

      public static final CalciteSystemProperty<Boolean> ENABLE_STREAM
      Whether streaming is enabled in the default planner configuration.

    • ENABLE_REX_DIGEST_NORMALIZE

      public static final CalciteSystemProperty<Boolean> ENABLE_REX_DIGEST_NORMALIZE
      Whether RexNode digest should be normalized (e.g. call operands ordered).

      Normalization helps to treat $0=$1 and $1=$0 expressions equal, thus it saves efforts on planning.

    • STRICT

      public static final CalciteSystemProperty<Boolean> STRICT
      Whether to follow the SQL standard strictly.

    • DUMP_GRAPHVIZ

      public static final CalciteSystemProperty<Boolean> DUMP_GRAPHVIZ
      Whether to include a GraphViz representation when dumping the state of the Volcano planner.

    • DUMP_SETS

      public static final CalciteSystemProperty<Boolean> DUMP_SETS
      Whether to include RelSet information when dumping the state of the Volcano planner.

    • TOPDOWN_OPT

      public static final CalciteSystemProperty<Boolean> TOPDOWN_OPT
      Whether to enable top-down optimization. This config can be overridden by CalciteConnectionProperty.TOPDOWN_OPT.

      Note: Enabling top-down optimization will automatically disable the use of AbstractConverter and related rules.

    • INTEGRATION_TEST

      public static final CalciteSystemProperty<Boolean> INTEGRATION_TEST
      Whether to run integration tests.

    • TEST_DB

      public static final CalciteSystemProperty<String> TEST_DB
      Which database to use for tests that require a JDBC data source.

      The property can take one of the following values:

      • HSQLDB (default)
      • H2
      • MYSQL
      • ORACLE
      • POSTGRESQL

      If the specified value is not included in the previous list, the default is used.

      We recommend that casual users use hsqldb, and frequent Calcite developers use MySQL. The test suite runs faster against the MySQL database (mainly because of the 0.1 second versus 6 seconds startup time). You have to populate MySQL manually with the foodmart data set, otherwise there will be test failures.

    • TEST_DATASET_PATH

      public static final CalciteSystemProperty<String> TEST_DATASET_PATH
      Path to the dataset file that should used for integration tests.

      If a path is not set, then one of the following values will be used:

      • ../calcite-test-dataset
      • ../../calcite-test-dataset
      • .
      The first valid path that exists in the filesystem will be chosen.

    • TEST_ARROW

      public static final CalciteSystemProperty<Boolean> TEST_ARROW
      Whether to run Arrow tests.

    • TEST_MONGODB

      public static final CalciteSystemProperty<Boolean> TEST_MONGODB
      Whether to run MongoDB tests.

    • TEST_SPLUNK

      public static final CalciteSystemProperty<Boolean> TEST_SPLUNK
      Whether to run Splunk tests.

      Disabled by default, because we do not expect Splunk to be installed and populated with the data set necessary for testing.

    • TEST_DRUID

      public static final CalciteSystemProperty<Boolean> TEST_DRUID
      Whether to run Druid tests.

    • TEST_CASSANDRA

      public static final CalciteSystemProperty<Boolean> TEST_CASSANDRA
      Whether to run Cassandra tests.

    • TEST_INNODB

      public static final CalciteSystemProperty<Boolean> TEST_INNODB
      Whether to run InnoDB tests.

    • TEST_REDIS

      public static final CalciteSystemProperty<Boolean> TEST_REDIS
      Whether to run Redis tests.

    • TEST_WITH_DOCKER_CONTAINER

      public static final CalciteSystemProperty<Boolean> TEST_WITH_DOCKER_CONTAINER
      Whether to use Docker containers in tests.

      If the property is set to true, affected tests will attempt to start Docker containers; when Docker is not available tests fallback to other execution modes and if it's not possible they are skipped entirely.

      If the property is set to false, Docker containers are not used at all and affected tests either fallback to other execution modes or skipped entirely.

      Users can override the default behavior to force non-Dockerized execution even when Docker is installed on the machine; this can be useful for replicating an issue that appears only in non-docker test mode or for running tests both with and without containers in CI.

    • TEST_FOODMART_QUERY_IDS

      public static final CalciteSystemProperty<@Nullable String> TEST_FOODMART_QUERY_IDS
      A list of ids designating the queries (from query.json in new.hydromatic:foodmart-queries:0.4.1) that should be run as part of FoodmartTest.

    • ALLOW_INFINITE_COST_CONVERTERS

      public static final CalciteSystemProperty<Boolean> ALLOW_INFINITE_COST_CONVERTERS
      Whether the optimizer will consider adding converters of infinite cost in order to convert a relational expression from one calling convention to another.

    • DEFAULT_CHARSET

      public static final CalciteSystemProperty<String> DEFAULT_CHARSET
      The name of the default character set.

      It is used by SqlValidator.

    • DEFAULT_NATIONAL_CHARSET

      public static final CalciteSystemProperty<String> DEFAULT_NATIONAL_CHARSET
      The name of the default national character set.

      It is used with the N'string' construct in SqlLiteral and may be different from the DEFAULT_CHARSET.

    • DEFAULT_COLLATION

      public static final CalciteSystemProperty<String> DEFAULT_COLLATION
      The name of the default collation.

      It is used in SqlCollation and SqlLiteral.

    • DEFAULT_COLLATION_STRENGTH

      public static final CalciteSystemProperty<String> DEFAULT_COLLATION_STRENGTH
      The strength of the default collation. Allowed values (as defined in Collator) are: primary, secondary, tertiary, identical.

      It is used in SqlCollation and SqlLiteral.

    • METADATA_HANDLER_CACHE_MAXIMUM_SIZE

      public static final CalciteSystemProperty<Integer> METADATA_HANDLER_CACHE_MAXIMUM_SIZE
      The maximum size of the cache of metadata handlers.

      A typical value is the number of queries being concurrently prepared multiplied by the number of types of metadata.

      If the value is less than 0, there is no limit.

    • BINDABLE_CACHE_MAX_SIZE

      public static final CalciteSystemProperty<Integer> BINDABLE_CACHE_MAX_SIZE
      The maximum size of the cache used for storing Bindable objects, instantiated via dynamically generated Java classes.

      The default value is 0.

      The property can take any value between [0, Integer.MAX_VALUE] inclusive. If the value is not valid (or not specified) then the default value is used.

      The cached objects may be quite big so it is suggested to use a rather small cache size (e.g., 1000). For the most common use cases a number close to 1000 should be enough to alleviate the performance penalty of compiling and loading classes.

      Setting this property to 0 disables the cache.

    • BINDABLE_CACHE_CONCURRENCY_LEVEL

      public static final CalciteSystemProperty<Integer> BINDABLE_CACHE_CONCURRENCY_LEVEL
      The concurrency level of the cache used for storing Bindable objects, instantiated via dynamically generated Java classes.

      The default value is 1.

      The property can take any value between [1, Integer.MAX_VALUE] inclusive. If the value is not valid (or not specified) then the default value is used.

      This property has no effect if the cache is disabled (i.e., BINDABLE_CACHE_MAX_SIZE set to 0.

    • FUNCTION_LEVEL_CACHE_MAX_SIZE

      public static final CalciteSystemProperty<Integer> FUNCTION_LEVEL_CACHE_MAX_SIZE
      The maximum number of items in a function-level cache.

      A few SQL functions have expensive processing that, if its results are cached, can be reused by future calls to the function. One such function is RLIKE, whose arguments are a regular expression and a string. The regular expression needs to be compiled to a Pattern. Compilation is expensive, and within a particular query, the arguments are often the same string, or a small number of distinct strings, so caching makes sense.

      Therefore, functions such as RLIKE, SIMILAR TO, PARSE_URL, PARSE_TIMESTAMP, FORMAT_DATE have a function-level cache. The cache is created in the code generated for the query, at the call site of the function, and expires when the query has finished executing. Such caches do not need time-based expiration, but we need to cap the size of the cache to deal with scenarios such as a billion-row table where every row has a distinct regular expression.

      Because of how Calcite generates and executes code in Enumerable convention, each function object is used from a single thread. Therefore, non thread-safe objects such as DateFormat can be safely cached.

      The value of this parameter limits the size of every function-level cache in Calcite. The default value is 1,000.

    • JOIN_SELECTOR_COMPACT_CODE_THRESHOLD

      public static final CalciteSystemProperty<Integer> JOIN_SELECTOR_COMPACT_CODE_THRESHOLD
      Minimum numbers of fields in a Join result that will trigger the "compact code generation". This feature reduces the risk of running into a compilation error due to the code of a dynamically generated method growing beyond the 64KB limit.

      Note that the compact code makes use of arraycopy operations when possible, instead of using a static array initialization. For joins with a large number of fields the resulting code should be faster, but it can be slower for joins with a very small number of fields.

      The default value is 100, a negative value disables completely the "compact code" feature.

      See Also:

  • Method Details

    • value

      public T value()
      Returns the value of this property.
      Returns:
      the value of this property or null if a default value has not been defined for this property.