oracle在执行一些DML操作时，会在block上有活动事务的标志，如果一个事务commit后，由于某些block在commit之前已经写回datafile, 或者事务影响到的block数过多，则commit的时候只会清理undo segment header中的事务标志信息，data block上的事务标志不会清除，否则代价过高。那么在一些读取这些block时，需要将这些事务标志进行清除，就是延迟块清除
实验如下：
SQL> drop table t cascade constraints;
Table dropped.
--创建一张表t
SQL> create table t as select * from dba_objects where 1=2;
Table created.
SQL> select count(*) from t;
COUNT(*)
----------
0
--采用直接路径插入的方式，即不经过buffer cache
SQL> insert /*+ append */ into t select * from dba_objects;
87023 rows created.
SQL> set autotrace on;
SQL> commit; --提交，此时新插入数据的block上的事务标志并没有清除
Commit complete.
SQL> select count(*) from t;
COUNT(*)
----------
87023
Execution Plan
----------------------------------------------------------
Plan hash value: 2966233522
-------------------------------------------------------------------
| Id | Operation | Name | Rows | Cost (%CPU)| Time |
-------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 1 | 339 (1)| 00:00:05 |
| 1 | SORT AGGREGATE | | 1 | | |
| 2 | TABLE ACCESS FULL| T | 67743 | 339 (1)| 00:00:05 |
-------------------------------------------------------------------
Note
-----
- dynamic sampling used for this statement (level=2)
Statistics
----------------------------------------------------------
0 recursive calls
1 db block gets
1249 consistent gets
1241 physical reads
168 redo size --查询即产生了redo，查询导致了data block上进行事务清除
528 bytes sent via SQL*Net to client
524 bytes received via SQL*Net from client
2 SQL*Net roundtrips to/from client
0 sorts (memory)
0 sorts (disk)
1 rows processed
SQL> /
COUNT(*)
----------
87023
Execution Plan
----------------------------------------------------------
Plan hash value: 2966233522
-------------------------------------------------------------------
| Id | Operation | Name | Rows | Cost (%CPU)| Time |
-------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 1 | 339 (1)| 00:00:05 |
| 1 | SORT AGGREGATE | | 1 | | |
| 2 | TABLE ACCESS FULL| T | 67743 | 339 (1)| 00:00:05 |
-------------------------------------------------------------------
Note
-----
- dynamic sampling used for this statement (level=2)
Statistics
----------------------------------------------------------
0 recursive calls
0 db block gets
1246 consistent gets
0 physical reads
0 redo size --再次查询不产生redo，事务已清除完毕
528 bytes sent via SQL*Net to client
524 bytes received via SQL*Net from client
2 SQL*Net roundtrips to/from client
0 sorts (memory)
0 sorts (disk)
1 rows processed
SQL> truncate table t;
Table truncated.
SQL> select count(*) from t;
COUNT(*)
----------
0
Execution Plan
----------------------------------------------------------
Plan hash value: 2966233522
-------------------------------------------------------------------
| Id | Operation | Name | Rows | Cost (%CPU)| Time |
-------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 1 | 2 (0)| 00:00:01 |
| 1 | SORT AGGREGATE | | 1 | | |
| 2 | TABLE ACCESS FULL| T | 1 | 2 (0)| 00:00:01 |
-------------------------------------------------------------------
Note
-----
- dynamic sampling used for this statement (level=2)
Statistics
----------------------------------------------------------
6 recursive calls
1 db block gets
13 consistent gets
0 physical reads
96 redo size --truncate表之后查询，同样也产生了延迟块清除
525 bytes sent via SQL*Net to client
524 bytes received via SQL*Net from client
2 SQL*Net roundtrips to/from client
0 sorts (memory)
0 sorts (disk)
1 rows processed