btree和點陣圖索引的對比
阿新 • • 發佈：2019-02-13
1、btree 索引 通過建表t1 object_id的值沒有重複值，而t2 表的object_id的值重複率很高 通過實驗在t1，t2表的object_id列建立普通索引，來證明普通索引列比較適合列的重複值比較低的列 優點：適合鍵值重複率較低的欄位上使用 那麼有個B-tree索引我們就像翻書目錄一樣，直接定位rowid立刻就找到了我們想要的資料，實質減少了I/O操作就提高速度，它有一 個顯著特點查詢效能與表中資料量無關 缺點：不適合鍵值重複率較高的欄位上使用， SQL> create table t1 as select object_id,object_name from dba_objects; Table created. SQL> create table t2 as select mod(object_id,2) object_id,object_name from dba_objects; Table created. SQL> create index ind_t1 on t1(object_id); Index created. SQL> create index ind_t2 on t2(object_id); Index created. 收集統計資訊： BEGIN DBMS_STATS.GATHER_TABLE_STATS(ownname => 'scott', tabname => 't1', estimate_percent =>100, method_opt => 'for all columns size 1', degree => 8, cascade=>TRUE ); END; BEGIN DBMS_STATS.GATHER_TABLE_STATS(ownname => 'scott', tabname => 't2', estimate_percent =>100, method_opt => 'for all columns size 1', degree => 8, cascade=>TRUE ); END; SQL> select count(*) from t1 where object_id=1; COUNT(*) ---------- 0 Execution Plan ---------------------------------------------------------- Plan hash value: 2587783732 ---------------------------------------------------------------------------- | Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time | ---------------------------------------------------------------------------- | 0 | SELECT STATEMENT | | 1 | 5 | 1 (0)| 00:00:01 | | 1 | SORT AGGREGATE | | 1 | 5 | | | |* 2 | INDEX RANGE SCAN| IND_T1 | 1 | 5 | 1 (0)| 00:00:01 | ---------------------------------------------------------------------------- Predicate Information (identified by operation id): --------------------------------------------------- 2 - access("OBJECT_ID"=1) Statistics ---------------------------------------------------------- 1 recursive calls 0 db block gets 2 consistent gets 0 physical reads 0 redo size 525 bytes sent via SQL*Net to client 523 bytes received via SQL*Net from client 2 SQL*Net roundtrips to/from client 0 sorts (memory) 0 sorts (disk) 1 rows processed SQL> select count(*) from t2 where object_id=1; COUNT(*) ---------- 36200 Execution Plan ---------------------------------------------------------- Plan hash value: 2800912005 -------------------------------------------------------------------------------- | Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time | -------------------------------------------------------------------------------- | 0 | SELECT STATEMENT | | 1 | 3 | 38 (0)| 00:00:01 | | 1 | SORT AGGREGATE | | 1 | 3 | | | |* 2 | INDEX FAST FULL SCAN| IND_T2 | 36086 | 105K| 38 (0)| 00:00:01 | -------------------------------------------------------------------------------- Predicate Information (identified by operation id): --------------------------------------------------- 2 - filter("OBJECT_ID"=1) Statistics ---------------------------------------------------------- 1 recursive calls 0 db block gets 144 consistent gets 0 physical reads 0 redo size 527 bytes sent via SQL*Net to client 523 bytes received via SQL*Net from client 2 SQL*Net roundtrips to/from client 0 sorts (memory) 0 sorts (disk) 1 rows processed SQL> drop index ind_t1; Index dropped. SQL> drop index ind_t2; Index dropped. 2、點陣圖索引 點陣圖索引適合於：列的基數很少，可列舉，重複值很多，資料不會被經常更新，由於一個鍵值對應很多行（rowid）， 更新索引鍵值的時候，就會鎖定索引，導致其他行不可被修改，阻塞 優點：OLAP 例如報表類資料庫 重複率高的資料 特定型別的查詢例如count、or、and等邏輯操作因為只需要進行位運算即可得到我們需要的結果 缺點：不適合重複率低的欄位，還有經常DML操作（insert，update，delete），因為點陣圖索引的鎖代價極高，修改一個位圖索引段影響整個點陣圖段，例如修改 一個鍵值，會影響同鍵值的多行，所以對於OLTP 系統點陣圖索引基本上是不適用的 接著上面的實驗，在t1 t2表上建立點陣圖索引 SQL> create bitmap index ind_t1 on t1(object_id); Index created. SQL> create bitmap index ind_t2 on t2(object_id); Index created. SQL> select segment_name,bytes from user_segments where segment_name like '%T1%' OR SEGMENT_NAME LIKE '%T2%'; SEGMENT_NAME BYTES --------------------------------------------------------------------------------- ---------- T1 3145728 T2 3145728 IND_T1 3145728 IND_T2 65536 我們可以看出t1表的object_id列沒有重複值，而t2表的object_id列重複值很多，建立點陣圖索引的時候，重複值越多，點陣圖索引就越小 SQL> drop table t1; SQL> drop table t2; 下面我們來看一下，在重複率很高的情況下，點陣圖索引和btree的效率 create table t1 as select object_id,object_type from dba_objects; create table t2 as select object_id,object_type from dba_objects; create index ind_t1 on t1(object_type); create bitmap index ind_t2 on t2(object_type); SQL> select count(*) from t1 where object_type='TABLE'; Execution Plan ---------------------------------------------------------- Plan hash value: 2587783732 ---------------------------------------------------------------------------- | Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time | ---------------------------------------------------------------------------- | 0 | SELECT STATEMENT | | 1 | 9 | 5 (0)| 00:00:01 | | 1 | SORT AGGREGATE | | 1 | 9 | | | |* 2 | INDEX RANGE SCAN| IND_T1 | 1678 | 15102 | 5 (0)| 00:00:01 | ---------------------------------------------------------------------------- Predicate Information (identified by operation id): --------------------------------------------------- 2 - access("OBJECT_TYPE"='TABLE') Statistics ---------------------------------------------------------- 1 recursive calls 0 db block gets 8 consistent gets 0 physical reads 0 redo size 527 bytes sent via SQL*Net to client 523 bytes received via SQL*Net from client 2 SQL*Net roundtrips to/from client 0 sorts (memory) 0 sorts (disk) 1 rows processed SQL> select count(*) from t2 where object_type='TABLE'; Execution Plan ---------------------------------------------------------- Plan hash value: 2032664525 -------------------------------------------------------------------------------------- | Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time | -------------------------------------------------------------------------------------- | 0 | SELECT STATEMENT | | 1 | 9 | 1 (0)| 00:00:01 | | 1 | SORT AGGREGATE | | 1 | 9 | | | | 2 | BITMAP CONVERSION COUNT | | 1678 | 15102 | 1 (0)| 00:00:01 | |* 3 | BITMAP INDEX SINGLE VALUE| IND_T2 | | | | | -------------------------------------------------------------------------------------- Predicate Information (identified by operation id): --------------------------------------------------- 3 - access("OBJECT_TYPE"='TABLE') Statistics ---------------------------------------------------------- 1 recursive calls 0 db block gets 2 consistent gets 0 physical reads 0 redo size 526 bytes sent via SQL*Net to client 523 bytes received via SQL*Net from client 2 SQL*Net roundtrips to/from client 0 sorts (memory) 0 sorts (disk) 1 rows processed SQL> select * from t1 where object_type='TABLE'; 2799 rows selected. Execution Plan ---------------------------------------------------------- Plan hash value: 634656657 -------------------------------------------------------------------------------------- | Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time | -------------------------------------------------------------------------------------- | 0 | SELECT STATEMENT | | 1678 | 23492 | 26 (0)| 00:00:01 | | 1 | TABLE ACCESS BY INDEX ROWID| T1 | 1678 | 23492 | 26 (0)| 00:00:01 | |* 2 | INDEX RANGE SCAN | IND_T1 | 1678 | | 5 (0)| 00:00:01 | -------------------------------------------------------------------------------------- Predicate Information (identified by operation id): --------------------------------------------------- 2 - access("OBJECT_TYPE"='TABLE') Statistics ---------------------------------------------------------- 1 recursive calls 0 db block gets 427 consistent gets 0 physical reads 0 redo size 79004 bytes sent via SQL*Net to client 2569 bytes received via SQL*Net from client 188 SQL*Net roundtrips to/from client 0 sorts (memory) 0 sorts (disk) 2799 rows processed SQL> select * from t2 where object_type='TABLE'; 2800 rows selected. Execution Plan ---------------------------------------------------------- Plan hash value: 2737179948 --------------------------------------------------------------------------------------- | Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time | --------------------------------------------------------------------------------------- | 0 | SELECT STATEMENT | | 1678 | 23492 | 47 (0)| 00:00:01 | | 1 | TABLE ACCESS BY INDEX ROWID | T2 | 1678 | 23492 | 47 (0)| 00:00:01 | | 2 | BITMAP CONVERSION TO ROWIDS| | | | | | |* 3 | BITMAP INDEX SINGLE VALUE | IND_T2 | | | | | --------------------------------------------------------------------------------------- Predicate Information (identified by operation id): --------------------------------------------------- 3 - access("OBJECT_TYPE"='TABLE') Statistics ---------------------------------------------------------- 1 recursive calls 0 db block gets 235 consistent gets 0 physical reads 0 redo size 79020 bytes sent via SQL*Net to client 2569 bytes received via SQL*Net from client 188 SQL*Net roundtrips to/from client 0 sorts (memory) 0 sorts (disk) 2800 rows processed 在等值查詢中我們可以看出點陣圖索引的效率依言高於B-tree索引
上面實驗參考了http://www.itpub.net/thread-1700144-1-1.htmlcreate table t1 as select object_id,mod(object_id,2) id, object_name, object_type From dba_objects; create table t2 as select object_id,mod(object_id,2) id, object_name, object_type From dba_objects; create index ind_type_t1 on t1(object_type); create bitmap index ind_type_t2 on t2(object_type); create index ind_object_id_t1 on t1(object_id); create bitmap index ind_object_id_t2 on t2(object_id); create index ind_id_t1 on t1(id); create bitmap index ind_id_t2 on t2(id); SQL> select * from t1 where object_id in (1,2,10,20,30,50,60,70,40); 8 rows selected. Execution Plan ---------------------------------------------------------- Plan hash value: 1020377091 ------------------------------------------------------------------------------------------------- | Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time | ------------------------------------------------------------------------------------------------- | 0 | SELECT STATEMENT | | 9 | 369 | 10 (0)| 00:00:01 | | 1 | INLIST ITERATOR | | | | | | | 2 | TABLE ACCESS BY INDEX ROWID| T1 | 9 | 369 | 10 (0)| 00:00:01 | |* 3 | INDEX RANGE SCAN | IND_OBJECT_ID_T1 | 9 | | 9 (0)| 00:00:01 | ------------------------------------------------------------------------------------------------- Predicate Information (identified by operation id): --------------------------------------------------- 3 - access("OBJECT_ID"=1 OR "OBJECT_ID"=2 OR "OBJECT_ID"=10 OR "OBJECT_ID"=20 OR "OBJECT_ID"=30 OR "OBJECT_ID"=40 OR "OBJECT_ID"=50 OR "OBJECT_ID"=60 OR "OBJECT_ID"=70) Statistics ---------------------------------------------------------- 1 recursive calls 0 db block gets 12 consistent gets 1 physical reads 0 redo size 980 bytes sent via SQL*Net to client 523 bytes received via SQL*Net from client 2 SQL*Net roundtrips to/from client 0 sorts (memory) 0 sorts (disk) 8 rows processed SQL> select * from t2 where object_id in (1,2,10,20,30,50,60,70,40); 8 rows selected. Execution Plan ---------------------------------------------------------- Plan hash value: 3310774432 -------------------------------------------------------------------------------------------------- | Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time | -------------------------------------------------------------------------------------------------- | 0 | SELECT STATEMENT | | 9 | 369 | 11 (0)| 00:00:01 | | 1 | INLIST ITERATOR | | | | | | | 2 | TABLE ACCESS BY INDEX ROWID | T2 | 9 | 369 | 11 (0)| 00:00:01 | | 3 | BITMAP CONVERSION TO ROWIDS| | | | | | |* 4 | BITMAP INDEX SINGLE VALUE | IND_OBJECT_ID_T2 | | | | | -------------------------------------------------------------------------------------------------- Predicate Information (identified by operation id): --------------------------------------------------- 4 - access("OBJECT_ID"=1 OR "OBJECT_ID"=2 OR "OBJECT_ID"=10 OR "OBJECT_ID"=20 OR "OBJECT_ID"=30 OR "OBJECT_ID"=40 OR "OBJECT_ID"=50 OR "OBJECT_ID"=60 OR "OBJECT_ID"=70) Statistics ---------------------------------------------------------- 1 recursive calls 0 db block gets 15 consistent gets 1 physical reads 0 redo size 980 bytes sent via SQL*Net to client 523 bytes received via SQL*Net from client 2 SQL*Net roundtrips to/from client 0 sorts (memory) 0 sorts (disk) 8 rows processed 結論：在使用or 的情況下，object_id 重複率比較低的情況下，還是btree效率高一些
SQL> select * From t2 where object_type in ('INDEX','CLUSTER'); 3805 rows selected. Execution Plan ---------------------------------------------------------- Plan hash value: 879057093 --------------------------------------------------------------------------------------------- | Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time | --------------------------------------------------------------------------------------------- | 0 | SELECT STATEMENT | | 3357 | 134K| 127 (0)| 00:00:02 | | 1 | INLIST ITERATOR | | | | | | | 2 | TABLE ACCESS BY INDEX ROWID | T2 | 3357 | 134K| 127 (0)| 00:00:02 | | 3 | BITMAP CONVERSION TO ROWIDS| | | | | | |* 4 | BITMAP INDEX SINGLE VALUE | IND_TYPE_T2 | | | | | --------------------------------------------------------------------------------------------- Predicate Information (identified by operation id): --------------------------------------------------- 4 - access("OBJECT_TYPE"='CLUSTER' OR "OBJECT_TYPE"='INDEX') Statistics ---------------------------------------------------------- 1 recursive calls 0 db block gets 349 consistent gets 1 physical reads 0 redo size 198248 bytes sent via SQL*Net to client 3306 bytes received via SQL*Net from client 255 SQL*Net roundtrips to/from client 0 sorts (memory) 0 sorts (disk) 3805 rows processed SQL> select * From t1 where object_type in ('INDEX','CLUSTER'); 3805 rows selected. Execution Plan ---------------------------------------------------------- Plan hash value: 918902357 -------------------------------------------------------------------------------------------- | Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time | -------------------------------------------------------------------------------------------- | 0 | SELECT STATEMENT | | 3357 | 134K| 87 (0)| 00:00:02 | | 1 | INLIST ITERATOR | | | | | | | 2 | TABLE ACCESS BY INDEX ROWID| T1 | 3357 | 134K| 87 (0)| 00:00:02 | |* 3 | INDEX RANGE SCAN | IND_TYPE_T1 | 3357 | | 11 (0)| 00:00:01 | -------------------------------------------------------------------------------------------- Predicate Information (identified by operation id): --------------------------------------------------- 3 - access("OBJECT_TYPE"='CLUSTER' OR "OBJECT_TYPE"='INDEX') Statistics ---------------------------------------------------------- 1 recursive calls 0 db block gets 610 consistent gets 0 physical reads 0 redo size 198248 bytes sent via SQL*Net to client 3306 bytes received via SQL*Net from client 255 SQL*Net roundtrips to/from client 0 sorts (memory) 0 sorts (disk) 3805 rows processed 結論：在使用or的情況下，object_type重複率比較高的表，還是點陣圖索引效率高一些