回复
openGauss内核分析（十）：数据库索引的创建过程
joytrian
发布于 2023-5-5 11:00
浏览
0收藏
数据库索引可以提高数据的访问速度，openGauss支持唯一索引、多字段索引、部分索引和表达式索引。行存表（ASTORE存储引擎）支持的索引类型：btree（行存表缺省值）、hash、gin、gist。行存表（USTORE存储引擎）支持的索引类型：ubtree。列存表支持的索引类型：Psort（列存表缺省值）、btree、gin。全局临时表不支持GIN索引和Gist索引。
如上一篇分析数据库表的创建过程，standard_ProcessUtility函数会根据nodeTag(parsetree)的值来确定sql的操作类型，create table一般都是进入T_CreateStmt分支调用CreateCommand函数。create index则进入T_IndexStmt分支调用DefineIndex函数。在调用DefineIndex前会首先执行函数transformIndexStmt，如果语句没有指定索引类型则会使用缺省值。
if (stmt->accessMethod == NULL) {
if (!isColStore) {
/* row store using btree index by default */
if (!RelationIsUstoreFormat(rel)) {
stmt->accessMethod = DEFAULT_INDEX_TYPE; // 行存表ASTORE存储引擎缺省值btree
} else {
stmt->accessMethod = DEFAULT_USTORE_INDEX_TYPE; // 行存表USTORE存储引擎缺省值ubtree
}
……
} else {
/* column store using psort index by default */
stmt->accessMethod = DEFAULT_CSTORE_INDEX_TYPE; // 列存表缺省值psort
}
}
#define DEFAULT_INDEX_TYPE "btree"
#define DEFAULT_HASH_INDEX_TYPE "hash"
#define DEFAULT_CSTORE_INDEX_TYPE "psort"
#define DEFAULT_GIST_INDEX_TYPE "gist"
#define CSTORE_BTREE_INDEX_TYPE "cbtree"
#define DEFAULT_GIN_INDEX_TYPE "gin"
#define CSTORE_GINBTREE_INDEX_TYPE "cgin"
#define DEFAULT_USTORE_INDEX_TYPE "ubtree"
普通表索引
DefineInde
DefineIndex为创建索引主入口函数。通常创建索引以Share锁锁定表，允许并发查询，但禁上对表进行修改。如果创建索引时指定关键字CONCURRENTLY以不阻塞DML的方式创建索引，即允许读取和更新表，以ShareUpdateExclusiveLock锁锁定表。
lockmode = concurrent ? ShareUpdateExclusiveLock : ShareLock;
rel = heap_open(relationId, lockmode);如果没有指定索引名，ChooseIndexName根据规则生成索引名：
/*
* Select name for index if caller didn't specify.
*/
indexRelationName = stmt->idxname;
if (indexRelationName == NULL) {
indexRelationName = ChooseIndexName(RelationGetRelationName(rel),
namespaceId,
indexColNames,
stmt->excludeOpNames,
stmt->primary,
stmt->isconstraint);为index_create函数构造参数IndexInfo结构体：
/*
* Prepare arguments for index_create, primarily an IndexInfo structure.
* Note that ii_Predicate must be in implicit-AND format.
*/
indexInfo = makeNode(IndexInfo);Index_create函数创建索引：
/*
* Make the catalog entries for the index, including constraints. Then, if
* not skip_build || concurrent, actually build the index.
*/
indexRelationId = index_create(rel,
……关闭表并返回索引表id：
heap_close(rel, NoLock);
return indexRelationId;
Index_create函数
打开系统表pg_class：
pg_class = heap_open(RelationRelationId, RowExclusiveLock);heap_create创建relcache和索引物理文件：
/*
* create the index relation's relcache entry and physical disk file. (If
* we fail further down, it's the smgr's responsibility to remove the disk
* file again.)
*/
StorageType storage_type = RelationGetStorageType(heapRelation);
indexRelation = heap_create(indexRelationName, namespaceId, tableSpaceId, indexRelationId, relFileNode,
RELATION_CREATE_BUCKET(heapRelation) ? heapRelation->rd_bucketoid : InvalidOid, indexTupDesc, relKind,
relpersistence, isLocalPart, false, shared_relation, mapped_relation, allow_system_table_mods,
REL_CMPRS_NOT_SUPPORT, (Datum)reloptions, heapRelation->rd_rel->relowner, skip_create_storage,
isUstore ? TAM_USTORE : TAM_HEAP, /* XXX: Index tables are by default HEAP Table Type */
relindexsplit, storage_type, extra->crossBucket, accessMethodObjectId);将索引表元信息存入系统表pg_class：
/*
* store index's pg_class entry
*/
InsertPgClassTuple(
pg_class, indexRelation, RelationGetRelid(indexRelation), (Datum)0, reloptions, relKind, NULL);
/* done with pg_class */
heap_close(pg_class, RowExclusiveLock);将索引表元信息存入系统表pg_index：
UpdateIndexRelation(indexRelationId,
heapRelationId,
indexInfo,
……Index_build建立索引：
} else if (extra && (!extra->isPartitionedIndex || extra->isGlobalPartitionedIndex)) {
/* support regular index or GLOBAL partition index */
index_build(heapRelation, NULL, indexRelation, NULL, indexInfo, isprimary, false, PARTITION_TYPE(extra));
}
index_build
index_build调用index_build_storage，如果创建的是btree索引最终调用btbuild，如果是hash索引最终调用hashbuild，如果是psort则最终调用psortbuild，更多索引访问方法的信息可查看系统表pg_am。
stats = index_build_storage(targetHeapRelation, targetIndexRelation, indexInfo);
static IndexBuildResult* index_build_storage(Relation heapRelation, Relation indexRelation, IndexInfo* indexInfo)
{
RegProcedure procedure = indexRelation->rd_am->ambuild;
Assert(RegProcedureIsValid(procedure));
IndexBuildResult* stats = (IndexBuildResult*)DatumGetPointer(OidFunctionCall3(
procedure, PointerGetDatum(heapRelation), PointerGetDatum(indexRelation), PointerGetDatum(indexInfo)));
Assert(PointerIsValid(stats));
if (RELPERSISTENCE_UNLOGGED == heapRelation->rd_rel->relpersistence) {
index_build_init_fork(heapRelation, indexRelation);
}
return stats;
}btree索引的procedure为btbuild。
#0 btbuild (fcinfo=0x7fb4f9c63920) at nbtree.cpp:63
#1 0x00000000011fc07d in OidFunctionCall3Coll (functionId=338, collation=0, arg1=140415366247440, arg2=140415366237480, arg3=140415402419864) at fmgr.cpp:1857
#2 0x0000000000c16b77 in index_build_storage (heapRelation=0x7fb50006b410, indexRelation=0x7fb500068d28, indexInfo=0x7fb5022ea698) at index.cpp:2475
#3 0x0000000000c18097 in index_build (heapRelation=0x7fb50006b410, heapPartition=0x0, indexRelation=0x7fb500068d28, indexPartition=0x0, indexInfo=0x7fb5022ea698, isPrimary=false, isreindex=false,
partitionType=INDEX_CREATE_NONE_PARTITION, parallel=true) at index.cpp:2834以上函数调用栈如下：
#0 index_build_storage
#1 index_build
#2 index_create
#3 DefineIndex
#4 standard_ProcessUtility
#5 gsaudit_ProcessUtility_hook
#6 pgaudit_ProcessUtility
#7 hypo_utility_hook
#8 ProcessUtility
#9 PortalRunUtility
#10 PortalRun
#11 exec_simple_query
#12 PostgresMain
分区表索引
创建普通表索引语法如下：
CREATE [ UNIQUE ] INDEX [ CONCURRENTLY ] [ [schema_name.]index_name ]
{ ON table_name [ USING method ] | [ USING method ] ON table_name }
({ { column_name | ( expression ) } [ COLLATE collation ] [ opclass ] [ ASC | DESC ] [ NULLS { FIRST | LAST } ] }[, ...] )
[ index_option ]
[ WHERE predicate ];创建分区表索引语法：
CREATE [ UNIQUE ] INDEX [ [schema_name.]index_name ]
{ ON table_name [ USING method ] | [ USING method ] ON table_name }
( {{ column_name | ( expression ) } [ COLLATE collation ] [ opclass ] [ ASC | DESC ] [ NULLS LAST ] }[, ...] )
[ LOCAL [ ( { PARTITION index_partition_name [ TABLESPACE index_partition_tablespace ] } [, ...] ) ] | GLOBAL ]
[ index_option ]两者执行流程基本一致，分区表索引在DefineIndex中是遍历每个分区调用partition_index_create。
文章转载自公众号：openGauss
分类
标签
已于2023-5-5 11:00:33修改
赞
收藏
回复
相关推荐
