All Variables

Complete list of session variables available via SET and SHOW.

General

Variable	Default	Description
search_path	"$user", public	Schema search order for unqualified names
timezone	Etc/UTC	Time zone for displaying and interpreting timestamps
session_timezone	(empty)	User-provided session timezone name (e.g., America/Los_Angeles)
adjust_timestamp_to_session_timezone	false	Adjust timezone-less timestamp conversions to the session timezone
datestyle	ISO, MDY	Display format for date and time values
intervalstyle	postgres	Display format for interval values
client_encoding	UTF8	Client character set encoding
server_encoding	UTF8	Server character set encoding (read-only)
server_version	18.1	Server version (read-only)
application_name	(empty)	Application name reported in logs
standard_conforming_strings	on	Causes '...' strings to treat backslashes literally

Transactions

Variable	Default	Description
default_transaction_isolation	repeatable read	Isolation level for new transactions
transaction_isolation	repeatable read	Current transaction's isolation level
default_transaction_read_only	off	Default read-only status of new transactions

SereneDB-Specific

Variable	Default	Description
sdb_write_conflict_policy	emit_error	Write conflict policy: emit_error, do_nothing (skip conflicted rows), or replace
sdb_read_your_own_writes	true	Whether queries see uncommitted writes from the current transaction

Query Execution

Variable	Default	Description
execution_threads	0	Number of threads for query execution. 0 means serial execution in the connection thread
join_order_algorithm	cost	Join ordering algorithm: cost (default), greedy, or syntactic
query_max_memory_per_node	0B	Maximum memory per query per host. 0 means unlimited
task_writer_count	4	Number of parallel table writer operators per task
task_partitioned_writer_count	4	Number of parallel table writer operators for partitioned writes
min_table_rows_for_parallel_join_build	1000	Minimum table rows to trigger parallel hash join build

Numeric Formatting

Variable	Default	Description
extra_float_digits	1	Number of extra digits displayed for floating-point values
bytea_output	hex	Output format for bytea: hex or escape
integer_datetimes	on	Whether datetimes are integer-based (read-only)

Expression Evaluation

Variable	Default	Description
expression.eval_simplified	false	Use the simplified expression evaluation path
expression.track_cpu_usage	false	Track CPU usage for individual expressions. Can be expensive for small batches
expression.max_array_size_in_reduce	100000	Maximum array size allowed in reduce()
expression.max_compiled_regexes	100	Maximum compiled regex patterns per function instance per thread
track_operator_cpu_usage	true	Track CPU usage for individual operator stages
legacy_cast	false	Use legacy CAST operator behavior
cast_match_struct_by_name	false	Match struct fields by name instead of position during CAST
adaptive_filter_reordering_enabled	true	Reorder conjunction inputs based on evaluation time
hash_adaptivity_enabled	true	Allow adaptive hash mode in GROUP BY
enable_expression_evaluation_cache	true	Enable caches in expression evaluation for performance

Spilling

Controls for spilling data to disk when memory is exceeded.

Variable	Default	Description
spill_enabled	false	Global spilling flag
aggregation_spill_enabled	true	Aggregation spilling (requires spill_enabled)
join_spill_enabled	true	Join spilling (requires spill_enabled)
order_by_spill_enabled	true	ORDER BY spilling (requires spill_enabled)
window_spill_enabled	true	Window function spilling (requires spill_enabled)
writer_spill_enabled	true	Table writer spilling (requires spill_enabled)
row_number_spill_enabled	true	ROW_NUMBER spilling (requires spill_enabled)
topn_row_number_spill_enabled	true	Top-N ROW_NUMBER spilling (requires spill_enabled)
local_merge_spill_enabled	false	Local merge spilling (requires spill_enabled)
max_spill_run_rows	12582912	Max rows per spill run (~128 MB). 0 means no limit
max_spill_bytes	107374182400	Max total spill bytes per query (default 100 GB). 0 means no limit
max_spill_level	1	Max recursive spill levels. -1 means no limit
max_spill_file_size	0	Max spill file size. 0 means no limit
spill_compression_codec	none	Compression codec for spilled data
spill_write_buffer_size	1048576	Spill write buffer size in bytes
spill_read_buffer_size	1048576	Spill read buffer size in bytes

Batch Sizing

Variable	Default	Description
preferred_output_batch_bytes	10485760	Preferred output batch size in bytes
preferred_output_batch_rows	1024	Preferred output batch row count (used when row size is unknown)
max_output_batch_rows	10000	Maximum output batch row count
max_output_buffer_size	33554432	Maximum task output buffer size in bytes
table_scan_getoutput_time_limit_ms	5000	Table scan time limit per getOutput() call in ms. 0 means no limit

Exchange & Partitioning

Variable	Default	Description
max_local_exchange_buffer_size	33554432	Max local exchange buffer size before backpressure
max_local_exchange_partition_count	4294967295	Max local exchange partitions
exchange.max_buffer_size	33554432	Max ExchangeQueue buffer size in bytes
merge_exchange.max_buffer_size	134217728	Max merge exchange buffer size in bytes
max_page_partitioning_buffer_size	33554432	Max PartitionedOutput buffer size in bytes

Aggregation

Variable	Default	Description
max_partial_aggregation_memory	16777216	Max memory for partial aggregation
max_extended_partial_aggregation_memory	67108864	Max extended memory for partial aggregation
abandon_partial_aggregation_min_rows	100000	Min rows before abandoning partial aggregation
abandon_partial_aggregation_min_pct	80	Min reduction percentage before abandoning partial aggregation

Query Tracing

Variable	Default	Description
query_trace_enabled	false	Enable query tracing
query_trace_dir	(empty)	Directory for query trace data
query_trace_node_id	(empty)	Plan node ID to trace
query_trace_max_bytes	0	Max trace bytes. 0 disables tracing

Debug

These settings are for debugging only and may impact performance.

Variable	Default	Description
debug.validate_output_from_operators	false	Validate output vectors of every operator
debug_disable_expression_with_peeling	false	Disable dictionary peeling optimization
debug_disable_common_sub_expressions	false	Disable common sub-expression caching
debug_disable_expression_with_memoization	false	Disable expression memoization
debug_disable_expression_with_lazy_inputs	false	Disable lazy input loading