- Added support for iterativeindex scans

- Added support for iterativesearch

- Added casts for arrays to`sparsevec`

- Improved cost estimation

- Improved performance of HNSW inserts and on-disk index builds

CREATETABLEitems (embedding vector(3), category_idint) PARTITION BY LIST(category_id);

*Unreleased*

With approximate indexes, queries with filtering can return less results (due to post-filtering).

Starting with 0.8.0, you can enable iterativeindex scans. If too few results from the initial index scan match the filters, the scan will resume until enough results are found (or it reaches`hnsw.max_search_tuples` or`ivfflat.max_probes`). This can significantly improve recall.

Starting with 0.8.0, you can enable iterativesearch. If too few results from the initial index scan match the filters, the scan will resume until enough results are found (or it reaches`hnsw.max_search_tuples` or`ivfflat.max_probes`). This can significantly improve recall.

There are two modes for iterativescans: strict and relaxed (due to the streaming nature of Postgres executor).

There are two modes for iterativesearch: strict and relaxed (due to the streaming nature of Postgres executor).

Strict ensures results are in the exact order by distance

SEThnsw.iterative_scan= strict_order;

SEThnsw.iterative_search= strict_order;

Relaxed allows results to be slightly out of order by distance, but provides better recall

SEThnsw.iterative_scan= relaxed_order;

SEThnsw.iterative_search= relaxed_order;

SETivfflat.iterative_scan= relaxed_order;

SETivfflat.iterative_search= relaxed_order;

Note: IVFFlat only supports relaxedordering for iterativescans

Note: IVFFlat only supports relaxedorder for iterativesearch

With relaxed ordering, you can use a[materialized CTE](https://www.postgresql.org/docs/current/queries-with.html#QUERIES-WITH-CTE-MATERIALIZATION) to get strict ordering

#defineMarkGUCPrefixReserved(x) EmitWarningsOnPlaceholders(x)

staticconststructconfig_enum_entryhnsw_iterative_scan_options[]= {

{"off",HNSW_ITERATIVE_SCAN_OFF, false},

{"relaxed_order",HNSW_ITERATIVE_SCAN_RELAXED, false},

{"strict_order",HNSW_ITERATIVE_SCAN_STRICT, false},

staticconststructconfig_enum_entryhnsw_iterative_search_options[]= {

{"off",HNSW_ITERATIVE_SEARCH_OFF, false},

{"relaxed_order",HNSW_ITERATIVE_SEARCH_RELAXED, false},

{"strict_order",HNSW_ITERATIVE_SEARCH_STRICT, false},

{NULL,0, false}

};

inthnsw_ef_search;

inthnsw_iterative_scan;

inthnsw_iterative_search;

inthnsw_max_search_tuples;

doublehnsw_search_mem_multiplier;

inthnsw_lock_tranche_id;

"Valid range is 1..1000.",&hnsw_ef_search,

HNSW_DEFAULT_EF_SEARCH,HNSW_MIN_EF_SEARCH,HNSW_MAX_EF_SEARCH,PGC_USERSET,0,NULL,NULL,NULL);

DefineCustomEnumVariable("hnsw.iterative_scan","Sets themode foriterativescans",

NULL,&hnsw_iterative_scan,

HNSW_ITERATIVE_SCAN_OFF,hnsw_iterative_scan_options,PGC_USERSET,0,NULL,NULL,NULL);

DefineCustomEnumVariable("hnsw.iterative_search","Sets the iterativesearch mode",

NULL,&hnsw_iterative_search,

HNSW_ITERATIVE_SEARCH_OFF,hnsw_iterative_search_options,PGC_USERSET,0,NULL,NULL,NULL);

DefineCustomIntVariable("hnsw.max_search_tuples","Sets the max number of candidates to visit for iterativescans",

DefineCustomIntVariable("hnsw.max_search_tuples","Sets the max number of candidates to visit for iterativesearch",

NULL,&hnsw_max_search_tuples,

20000,1,INT_MAX,PGC_USERSET,0,NULL,NULL,NULL);

DefineCustomRealVariable("hnsw.search_mem_multiplier","Sets the multiple of work_mem to use for iterativescans",

DefineCustomRealVariable("hnsw.search_mem_multiplier","Sets the multiple of work_mem to use for iterativesearch",

NULL,&hnsw_search_mem_multiplier,

2,1,1000,PGC_USERSET,0,NULL,NULL,NULL);

externinthnsw_ef_search;

externinthnsw_iterative_scan;

externinthnsw_iterative_search;

externinthnsw_max_search_tuples;

externdoublehnsw_search_mem_multiplier;

externinthnsw_lock_tranche_id;

{

HNSW_ITERATIVE_SCAN_OFF,

HNSW_ITERATIVE_SCAN_RELAXED,

}HnswIterativeScanMode;

HNSW_ITERATIVE_SEARCH_OFF,

HNSW_ITERATIVE_SEARCH_RELAXED,

}HnswIterativeSearchMode;

typedefstructHnswElementDataHnswElementData;

typedefstructHnswNeighborArrayHnswNeighborArray;

ep=w;

}

returnHnswSearchLayer(base,q,ep,hnsw_ef_search,0,index,support,m, false,NULL,&so->v,hnsw_iterative_scan!=HNSW_ITERATIVE_SCAN_OFF ?&so->discarded :NULL, true,&so->tuples);

returnHnswSearchLayer(base,q,ep,hnsw_ef_search,0,index,support,m, false,NULL,&so->v,hnsw_iterative_search!=HNSW_ITERATIVE_SEARCH_OFF ?&so->discarded :NULL, true,&so->tuples);

}

if (list_length(so->w)==0)

{

if (hnsw_iterative_scan==HNSW_ITERATIVE_SCAN_OFF)

if (hnsw_iterative_search==HNSW_ITERATIVE_SEARCH_OFF)

break;

so->w=list_delete_last(so->w);

if (hnsw_iterative_scan!=HNSW_ITERATIVE_SCAN_OFF)

if (hnsw_iterative_search!=HNSW_ITERATIVE_SEARCH_OFF)

{

pfree(element);

pfree(sc);

heaptid=&element->heaptids[--element->heaptidsLength];

if (hnsw_iterative_scan==HNSW_ITERATIVE_SCAN_STRICT)

if (hnsw_iterative_search==HNSW_ITERATIVE_SEARCH_STRICT)

{

if (sc->distance<so->previousDistance)

continue;

intivfflat_probes;

intivfflat_iterative_scan;

intivfflat_iterative_search;

intivfflat_max_probes;

staticrelopt_kindivfflat_relopt_kind;

staticconststructconfig_enum_entryivfflat_iterative_scan_options[]= {

{"off",IVFFLAT_ITERATIVE_SCAN_OFF, false},

{"relaxed_order",IVFFLAT_ITERATIVE_SCAN_RELAXED, false},

staticconststructconfig_enum_entryivfflat_iterative_search_options[]= {

{"off",IVFFLAT_ITERATIVE_SEARCH_OFF, false},

{"relaxed_order",IVFFLAT_ITERATIVE_SEARCH_RELAXED, false},

{NULL,0, false}

};

"Valid range is 1..lists.",&ivfflat_probes,

IVFFLAT_DEFAULT_PROBES,IVFFLAT_MIN_LISTS,IVFFLAT_MAX_LISTS,PGC_USERSET,0,NULL,NULL,NULL);

DefineCustomEnumVariable("ivfflat.iterative_scan","Sets themode foriterativescans",

NULL,&ivfflat_iterative_scan,

IVFFLAT_ITERATIVE_SCAN_OFF,ivfflat_iterative_scan_options,PGC_USERSET,0,NULL,NULL,NULL);

DefineCustomEnumVariable("ivfflat.iterative_search","Sets the iterativesearch mode",

NULL,&ivfflat_iterative_search,

IVFFLAT_ITERATIVE_SEARCH_OFF,ivfflat_iterative_search_options,PGC_USERSET,0,NULL,NULL,NULL);

DefineCustomIntVariable("ivfflat.max_probes","Sets the max number of probes for iterativescans",

DefineCustomIntVariable("ivfflat.max_probes","Sets the max number of probes for iterativesearch",

NULL,&ivfflat_max_probes,

IVFFLAT_MAX_LISTS,IVFFLAT_MIN_LISTS,IVFFLAT_MAX_LISTS,PGC_USERSET,0,NULL,NULL,NULL);

externintivfflat_probes;

externintivfflat_iterative_scan;

externintivfflat_iterative_search;

externintivfflat_max_probes;

{

IVFFLAT_ITERATIVE_SCAN_OFF,

}IvfflatIterativeScanMode;

IVFFLAT_ITERATIVE_SEARCH_OFF,

}IvfflatIterativeSearchMode;

{

}

}

if (tuples<100&&ivfflat_iterative_scan==IVFFLAT_ITERATIVE_SCAN_OFF)

if (tuples<100&&ivfflat_iterative_search==IVFFLAT_ITERATIVE_SEARCH_OFF)

ereport(DEBUG1,

(errmsg("index scan found few tuples"),

errdetail("Index may have been created with little data."),

IvfflatGetMetaPageInfo(index,&lists,&dimensions);

if (ivfflat_iterative_scan!=IVFFLAT_ITERATIVE_SCAN_OFF)

if (ivfflat_iterative_search!=IVFFLAT_ITERATIVE_SEARCH_OFF)

maxProbes=Max(ivfflat_max_probes,probes);

maxProbes=probes;

CREATE TABLE t (val vector(3));

INSERT INTO t (val) VALUES ('[0,0,0]'), ('[1,2,3]'), ('[1,1,1]'), (NULL);

CREATE INDEX ON t USING hnsw (val vector_l2_ops);

SET hnsw.iterative_scan = strict_order;

SET hnsw.iterative_search = strict_order;

SET hnsw.ef_search = 1;

SELECT * FROM t ORDER BY val <-> '[3,3,3]';

val

[0,0,0]

(3 rows)

SET hnsw.iterative_scan = relaxed_order;

SET hnsw.iterative_search = relaxed_order;

SELECT * FROM t ORDER BY val <-> '[3,3,3]';

val

---------

[0,0,0]

(3 rows)

RESET hnsw.iterative_scan;

RESET hnsw.iterative_search;

RESET hnsw.ef_search;

DROP TABLE t;

-- unlogged

ERROR: 0 is outside the valid range for parameter "hnsw.ef_search" (1 .. 1000)

SET hnsw.ef_search = 1001;

ERROR: 1001 is outside the valid range for parameter "hnsw.ef_search" (1 .. 1000)

SHOW hnsw.iterative_scan;

hnsw.iterative_scan

---------------------

SHOW hnsw.iterative_search;

hnsw.iterative_search

-----------------------

off

(1 row)

SET hnsw.iterative_scan = on;

ERROR: invalid value for parameter "hnsw.iterative_scan": "on"

SET hnsw.iterative_search = on;

ERROR: invalid value for parameter "hnsw.iterative_search": "on"

HINT: Available values: off, relaxed_order, strict_order.

SHOW hnsw.max_search_tuples;

hnsw.max_search_tuples

CREATE TABLE t (val vector(3));

INSERT INTO t (val) VALUES ('[0,0,0]'), ('[1,2,3]'), ('[1,1,1]'), (NULL);

CREATE INDEX ON t USING ivfflat (val vector_l2_ops) WITH (lists = 3);

SET ivfflat.iterative_scan = relaxed_order;

SET ivfflat.iterative_search = relaxed_order;

SELECT * FROM t ORDER BY val <-> '[3,3,3]';

val

---------

[1,1,1]

(2 rows)

RESET ivfflat.iterative_scan;

RESET ivfflat.iterative_search;

RESET ivfflat.max_probes;

DROP TABLE t;

-- unlogged

ERROR: 0 is outside the valid range for parameter "ivfflat.probes" (1 .. 32768)

SET ivfflat.probes = 32769;

ERROR: 32769 is outside the valid range for parameter "ivfflat.probes" (1 .. 32768)

SHOW ivfflat.iterative_scan;

ivfflat.iterative_scan

------------------------

SHOW ivfflat.iterative_search;

ivfflat.iterative_search

--------------------------

off

(1 row)

SET ivfflat.iterative_scan = on;

ERROR: invalid value for parameter "ivfflat.iterative_scan": "on"

SET ivfflat.iterative_search = on;

ERROR: invalid value for parameter "ivfflat.iterative_search": "on"

HINT: Available values: off, relaxed_order.

SHOW ivfflat.max_probes;

ivfflat.max_probes

INSERT INTO t (val)VALUES ('[0,0,0]'), ('[1,2,3]'), ('[1,1,1]'), (NULL);

CREATEINDEXON t USING hnsw (val vector_l2_ops);

SEThnsw.iterative_scan= strict_order;

SEThnsw.iterative_search= strict_order;

SEThnsw.ef_search=1;

SELECT*FROM tORDER BY val<->'[3,3,3]';

SEThnsw.iterative_scan= relaxed_order;

SEThnsw.iterative_search= relaxed_order;

SELECT*FROM tORDER BY val<->'[3,3,3]';

RESEThnsw.iterative_scan;

RESEThnsw.iterative_search;

RESEThnsw.ef_search;

DROPTABLE t;

SEThnsw.ef_search=0;

SEThnsw.ef_search=1001;

SHOWhnsw.iterative_scan;

SHOWhnsw.iterative_search;

SEThnsw.iterative_scan=on;

SEThnsw.iterative_search=on;

SHOWhnsw.max_search_tuples;

INSERT INTO t (val)VALUES ('[0,0,0]'), ('[1,2,3]'), ('[1,1,1]'), (NULL);

CREATEINDEXON t USING ivfflat (val vector_l2_ops) WITH (lists=3);

SETivfflat.iterative_scan= relaxed_order;

SETivfflat.iterative_search= relaxed_order;

SELECT*FROM tORDER BY val<->'[3,3,3]';

SETivfflat.max_probes=1;

SETivfflat.max_probes=2;

SELECT*FROM tORDER BY val<->'[3,3,3]';

RESETivfflat.iterative_scan;

RESETivfflat.iterative_search;

RESETivfflat.max_probes;

DROPTABLE t;

SETivfflat.probes=0;

SETivfflat.probes=32769;

SHOWivfflat.iterative_scan;

SHOWivfflat.iterative_search;

SETivfflat.iterative_scan=on;

SETivfflat.iterative_search=on;

SHOWivfflat.max_probes;

my$count =$node->safe_psql("postgres",qq(

));

is($count, 10);

$count =$node->safe_psql("postgres",qq(

));