解决direct path read 与 direct path write问题

1.这个问题主要与sort有关。

2.确定对象的sql：

SELECT segment_name, partition_name, p1, p2, wait1.sql_hash_value

FROM dba_extents, wait1

WHERE wait1.p2 BETWEEN block_id AND (block_id + blocks - 1)

AND file_id = wait1.p1

AND wait1.event = 'direct path read'

ORDER BY segment_name

3.根据回话读取的数据库文件类型了解回话所做的事情。如果是临时文件，则回话正在读取先前通过direct path write操作的临时段。如果是数据文件，可能是并行查询从属操作在工作。

SELECT NAME

FROM v$datafile

WHERE file# = :p1

UNION ALL

SELECT a.NAME

FROM v$tempfile a, v$parameter b

WHERE b.NAME = 'db_files' AND a.file# + b.VALUE = :p1

4.如果回话读取临时段，要查明是什么类型的段：

SELECT DISTINCT DECODE (ktssosegt,

1, 'SORT',

2, 'HASH',

3, 'DATA',

4, 'INDEX',

5, 'LOB_DATA',

6, 'LOB_INDEX',

'UNDEFINED'

)

FROM SYS.x$ktsso

WHERE inst_id = USERENV ('instance')

AND ktssoses = :cursor_session_address

AND ktssosno = :cursor_serial# ；

5.获取sql语句：

SELECT hash_value, address, piece, sql_text

FROM v$sqltext

WHERE hash_value = :1

ORDER BY hash_value, piece;