什么是集群因子（Clustering Factor）呢？集群因子是通過(guò)一個(gè)索引掃描一張表時(shí)需要訪問(wèn)的表的數(shù)據(jù)塊的數(shù)量。

集群因子的計(jì)算方法如下：

(1) 掃描一個(gè)索引；

(2) 比較某行的ROWID和前一行的ROWID，如果這兩個(gè)ROWID不屬于同一個(gè)數(shù)據(jù)塊，那么集群因子增加1；

(3) 整個(gè)索引掃描完畢后，就得到了該索引的集群因子。

以上面集群因子的計(jì)算方式可以看出，集群因子反映了索引范圍掃描可能帶來(lái)的對(duì)整個(gè)表訪問(wèn)過(guò)程的IO開(kāi)銷情況，如果集群因子接近于表存儲(chǔ)的塊數(shù)，說(shuō)明這張表是按照索引字段的順序存儲(chǔ)的。如果集群因子接近于行的數(shù)量，那說(shuō)明這張表不是按索引字段順序存儲(chǔ)的。在計(jì)算索引訪問(wèn)成本時(shí)，集群因子十分有用。集群因子乘以選擇性參數(shù)就是訪問(wèn)索引的開(kāi)銷。（是優(yōu)化器決定是否用索引的關(guān)鍵因素）

官方說(shuō)明：

?----row存儲(chǔ)的越有序，clustering factor的值越低

----當(dāng)clustering factor很高時(shí)，說(shuō)明index entry (rowid) 是????隨機(jī)指向一些block的，在一個(gè)大的index range scan時(shí)，這樣為了讀取這些rowid指向的block，就需要一次又一次重復(fù)的去讀這些block。

? ? ? ?----當(dāng)clustering factor值低時(shí)，說(shuō)明index keys (rowid) 是指向的記錄是存儲(chǔ)在相同的block里，這樣去讀row時(shí)，只需要在同一個(gè)block里讀取就可以了，這樣減少重復(fù)讀取blocks的次數(shù)

Index Clustering Factor說(shuō)明

??? 簡(jiǎn)單的說(shuō)，Index Clustering Factor是通過(guò)一個(gè)索引掃描一張表，需要訪問(wèn)的表的數(shù)據(jù)塊的數(shù)量，即對(duì)I/O的影響，也代表索引鍵存儲(chǔ)位置是否有序。

????(1)、如果越有序，即相鄰的鍵值存儲(chǔ)在相同的block，那么這時(shí)候Clustering Factor的值就越低；

??? (2)、如果不是很有序，即鍵值是隨機(jī)的存儲(chǔ)在block上，這樣在讀取鍵值時(shí)，可能就需要一次又一次的去訪問(wèn)相同的block，從而增加了I/O。

??? Clustering Factor的計(jì)算方式如下：

???? (1)、掃描一個(gè)索引(large index range scan)；

???? (2)、比較某行的rowid和前一行的rowid，如果這兩個(gè)rowid不屬于同一個(gè)數(shù)據(jù)塊，那么cluster factor增加1；

???? (3)、整個(gè)索引掃描完畢后，就得到了該索引的clustering factor。

??????????? 如果clustering factor接近于表存儲(chǔ)的塊數(shù)，說(shuō)明這張表是按照索引字段的順序存儲(chǔ)的。

??????????? 如果clustering factor接近于行的數(shù)量，那說(shuō)明這張表不是按索引字段順序存儲(chǔ)的。

??????????? 在計(jì)算索引訪問(wèn)成本的時(shí)候，這個(gè)值十分有用。Clustering Factor乘以選擇性參數(shù)(selectivity)就是訪問(wèn)索引的開(kāi)銷。

??????????? 如果這個(gè)統(tǒng)計(jì)數(shù)據(jù)不能真實(shí)反映出索引的真實(shí)情況，那么可能會(huì)造成優(yōu)化器錯(cuò)誤的選擇執(zhí)行計(jì)劃。另外如果某張表上的大多數(shù)訪問(wèn)是按照某個(gè)索引做索引掃描，那么將該表的數(shù)據(jù)按照索引字段的順序重新組織，可以提高該表的訪問(wèn)性能。

測(cè)試

????產(chǎn)生問(wèn)題

----查看一下數(shù)據(jù)庫(kù)的版本---- SQL>?select?*?from?v$version?where?rownum=1;BANNER -------------------------------------------------------------------------------- Oracle?Database?11g?Enterprise?Edition?Release?11.2.0.1.0?-?Production----創(chuàng)建一張測(cè)試表jack---- SQL>?create?table?jack?as?select?*?from?dba_objects?where?1=2;Table?created.----將數(shù)據(jù)無(wú)序的插入jack表中---- SQL>?begin2??????for?i?in?1..10?loop3????????insert?/*+?append?*/?into?jack?select?*?from?dba_objects?order?by?i;4??????commit;5????end?loop;6??end;7??/PL/SQL?procedure?successfully?completed.SQL>?select?count(*)?from?jack;COUNT(*) ----------725460----查看一下表的大小----- SQL>?set?wrap?off SQL>?col?owner?for?a10; SQL>?col?segment_name?for?a15; SQL>?select?segment_name,blocks,extents,bytes/1024/1024||'M'?"size"?from?user_segments?where?segment_name='JACK';SEGMENT_NAME?????BLOCKS????EXTENTS???size -------------?----------?----------?-------- JACK?????????????11264???????82??????88M----在object_id上創(chuàng)建索引---- SQL>?create?index?jack_ind?on?jack(object_id);Index?created.----查看一下索引的大小---- SQL>?select?segment_name,segment_type,blocks,extents,bytes/1024/1024||'M'?"size"?from?user_segments?where?segment_name='JACK_IND';SEGMENT_NAME????SEGMENT_TYPE??????BLOCKS?????EXTENTS?????size ------------?------------------?----------?----------?--------- JACK_IND???????????INDEX???????????1664?????????28????????13M----在沒(méi)有收集相關(guān)的統(tǒng)計(jì)信息之前，查看一下index?clustering?factor---- SQL>?select?index_name,clustering_factor,num_rows?from?user_indexes?where?index_name='JACK_IND';INDEX_NAME??????CLUSTERING_FACTOR???NUM_ROWS ---------------?-----------------?---------- JACK_IND??????????????725460?????????725460----簡(jiǎn)單的收集一下統(tǒng)計(jì)信息---- SQL>?exec?dbms_stats.gather_table_stats(user,'jack',cascade=>true);PL/SQL?procedure?successfully?completed.----再次查看index?clustering?factor---- SQL>?select?index_name,clustering_factor,num_rows?from?user_indexes?where?index_name='JACK_IND';INDEX_NAME?????CLUSTERING_FACTOR???NUM_ROWS --------------?-----------------?---------- JACK_IND??????????????725460???????725460???????----顯然統(tǒng)計(jì)信息收集前和后，clustering?factor值不變，說(shuō)在創(chuàng)建索引的時(shí)候，會(huì)收集表中的數(shù)據(jù)真正的行數(shù)。并且這里的clustering?factor等num_rows，也說(shuō)明表的clustering?factor是無(wú)序的。----查看一個(gè)確定值，然后查看執(zhí)行計(jì)劃---- SQL>?explain?plan?for?select?*?from?jack?where?object_id=1501;Explained.SQL>?select?*?from?table(dbms_xplan.display);PLAN_TABLE_OUTPUT -------------------------------------------------------------------------------- Plan?hash?value:?2860868395-------------------------------------------------------------------------------- |?Id??|?Operation????????????|?Name?????|?Rows??|?Bytes?|?Cost?(%CPU)|?Ti -------------------------------------------------------------------------------- |???0?|?SELECT?STATEMENT????????|???????????|????10?|???970?|????13?????(0)|?00 |???1?|??TABLE?ACCESS?BY?INDEX?ROWID|?JACK?????|????10?|???970?|????13?????(0)|?00 |*??2?|???INDEX?RANGE?SCAN????????|?JACK_IND?|????10?|???????|?????3?????(0)|?00 --------------------------------------------------------------------------------Predicate?Information?(identified?by?operation?id):PLAN_TABLE_OUTPUT --------------------------------------------------------------------------------2?-?access("OBJECT_ID"=1501)14?rows?selected.?????????????????----在這里走了索引，cost為13.SQL>?alter?system?flush?buffer_cache;System?altered.SQL>?set?autotrace?traceonly;----查詢一個(gè)范圍的執(zhí)行計(jì)劃---- SQL>?select?*?from?jack?where?object_id>1000?and?object_id<2000;9880?rows?selected.Execution?Plan ---------------------------------------------------------- Plan?hash?value:?949574992-------------------------------------------------------------------------- |?Id??|?Operation??????|?Name?|?Rows??|?Bytes?|?Cost?(%CPU)|?Time?????| -------------------------------------------------------------------------- |???0?|?SELECT?STATEMENT??|?????|??9657?|???914K|??1824???(1)|?00:00:22?| |*??1?|??TABLE?ACCESS?FULL|?JACK?|??9657?|???914K|??1824???(1)|?00:00:22?| --------------------------------------------------------------------------Predicate?Information?(identified?by?operation?id): ---------------------------------------------------1?-?filter("OBJECT_ID"<2000?AND?"OBJECT_ID">1000)Statistics ----------------------------------------------------------0???recursive?calls0???db?block?gets10993?consistent?gets10340?physical?reads0???redo?size471945?bytes?sent?via?SQL*Net?to?client7657??bytes?received?via?SQL*Net?from?client660???SQL*Net?roundtrips?to/from?client0???sorts?(memory)0???sorts?(disk)9880??rows?processed???????----注意，object_id上是有索引的，但是這里并沒(méi)有使用索引，而是使用了全表掃描。SQL>?alter?system?flush?buffer_cache;System?altered.----強(qiáng)制走索引，查看執(zhí)行計(jì)劃---- SQL>?select?/*+?index(jack?jack_ind)?*/?*?from?jack?where?object_id>1000?and?object_id<2000;9880?rows?selected.Execution?Plan ---------------------------------------------------------- Plan?hash?value:?2860868395---------------------------------------------------------------------------------------- |?Id??|?Operation????????????|?Name?????|?Rows??|?Bytes?|?Cost?(%CPU)|?Time?????| ---------------------------------------------------------------------------------------- |???0?|?SELECT?STATEMENT????????|???????????|??9657?|???914K|??9683?????(1)|?00:01:57?| |???1?|??TABLE?ACCESS?BY?INDEX?ROWID|?JACK?????|??9657?|???914K|??9683?????(1)|?00:01:57?| |*??2?|???INDEX?RANGE?SCAN????????|?JACK_IND?|??9657?|???????|????24?????(0)|?00:00:01?| ----------------------------------------------------------------------------------------Predicate?Information?(identified?by?operation?id): ---------------------------------------------------2?-?access("OBJECT_ID">1000?AND?"OBJECT_ID"<2000)Statistics ----------------------------------------------------------0????recursive?calls0????db?block?gets10561??consistent?gets164????physical?reads0????redo?size988947??bytes?sent?via?SQL*Net?to?client7657???bytes?received?via?SQL*Net?from?client660????SQL*Net?roundtrips?to/from?client0????sorts?(memory)0????sorts?(disk)9880???rows?processed???? ----強(qiáng)制走索引之后，使用了index?range?scan,但是cost變成了9683，而全表掃描時(shí)是1824. ----還有比較一下兩次查詢中物理讀的情況：全表掃描的物理讀明顯比索引的要高很多，但是Oracle卻沒(méi)有使用索引。 ----因此Oracle認(rèn)為走索引的Cost比走全表掃描大，而是大N倍，CBO是基于Cost來(lái)決定執(zhí)行計(jì)劃的。 ----由此得出，對(duì)于索引的Cost，Oracle是根據(jù)clustering?factor參數(shù)來(lái)計(jì)算的，而該實(shí)驗(yàn)中的clustering?factor參數(shù)是很高的，數(shù)據(jù)存儲(chǔ)無(wú)序。這就造成了Oracle認(rèn)為走索引的cost比全表掃描的大。

解決問(wèn)題：

----通過(guò)上面的分析，可以看出，要降低clustering?factor才能解決問(wèn)題，而要解決clustering?factor，就需要重新對(duì)表的存儲(chǔ)位置進(jìn)行排序。--------重建jakc表---- SQL>?create?table?echo?as?select?*?from?jack?where?1=0;Table?created.SQL>?insert?/*+?append?*/?into?echo?select?*?from?jack?order?by?object_id;725460?rows?created.SQL>?commit;Commit?complete.SQL>?truncate?table?jack;Table?truncated.SQL>?insert?/*+?append?*/?into?jack?select?*?from?echo;725460?rows?created.SQL>?commit;Commit?complete.----查看表和索引的信息---- SQL>?select?segment_name,blocks,extents,bytes/1024/1024||'M'?"size"?from?user_segments?where?segment_name='JACK';SEGMENT_NAME????BLOCKS????EXTENTS?????size -------------?----------?----------?----------- JACK?????????????11264???????82????????88MSQL>?select?segment_name,segment_type,blocks,extents,bytes/1024/1024||'M'?"size"?from?user_segments?where?segment_name='JACK_IND';SEGMENT_NAME????SEGMENT_TYPE??????BLOCKS?????EXTENTS????size ------------?------------------?----------?----------?------------- JACK_IND????????????INDEX??????????1536??????????27?????12MSQL>?select?index_name,clustering_factor,num_rows?from?user_indexes?where?index_name='JACK_IND';INDEX_NAME?????CLUSTERING_FACTOR?NUM_ROWS -------------?-----------------?---------- JACK_IND?????????????725460??????725460----對(duì)索引進(jìn)行rebuild---- SQL>?alter?index?jack_ind?rebuild;Index?altered.----查看cluster?factor---- SQL>?select?index_name,clustering_factor,num_rows?from?user_indexes?where?index_name='JACK_IND';INDEX_NAME??????CLUSTERING_FACTOR??NUM_ROWS ---------------?-----------------?---------- JACK_IND???????????????10327???????725460????------注意這里的Factor，已經(jīng)變成10327，我們收集一下表的統(tǒng)計(jì)信息，然后與表的block進(jìn)行一次比較。SQL>?exec?dbms_stats.gather_table_stats(user,'jack',cascade=>true);PL/SQL?procedure?successfully?completed.SQL>?select?blocks?from?dba_tables?where?table_name='JACK';BLOCKS ----------10474???----表jack實(shí)際使用的block是10474，clustering?factor是10327基本還是比較接近了，這也說(shuō)明相鄰的row是存儲(chǔ)在相同的block里。SQL>?select?index_name,clustering_factor,num_rows?from?user_indexes?where?index_name='JACK_IND';INDEX_NAME???????????????CLUSTERING_FACTOR???NUM_ROWS ------------------------------?-----------------?---------- JACK_IND???????????????????10327?????725460SQL>?alter?system?flush?buffer_cache;System?altered.SQL>?set?autotrace?traceonly;----再次查看之前sql的執(zhí)行計(jì)劃---- SQL>?select?*?from?jack?where?object_id>1000?and?object_id<2000;9880?rows?selected.Execution?Plan ---------------------------------------------------------- Plan?hash?value:?2860868395---------------------------------------------------------------------------------------- |?Id??|?Operation????????????|?Name?????|?Rows??|?Bytes?|?Cost?(%CPU)|?Time?????| ---------------------------------------------------------------------------------------- |???0?|?SELECT?STATEMENT????????|???????????|??9657?|???914K|???162?????(0)|?00:00:02?| |???1?|??TABLE?ACCESS?BY?INDEX?ROWID|?JACK?????|??9657?|???914K|???162?????(0)|?00:00:02?| |*??2?|???INDEX?RANGE?SCAN????????|?JACK_IND?|??9657?|???????|????24?????(0)|?00:00:01?| ----------------------------------------------------------------------------------------Predicate?Information?(identified?by?operation?id): ---------------------------------------------------2?-?access("OBJECT_ID">1000?AND?"OBJECT_ID"<2000)Statistics ----------------------------------------------------------1???recursive?calls0???db?block?gets1457??consistent?gets151???physical?reads0???redo?size988947??bytes?sent?via?SQL*Net?to?client7657??bytes?received?via?SQL*Net?from?client660???SQL*Net?roundtrips?to/from?client0???sorts?(memory)0???sorts?(disk)9880??rows?processed----注意這里的cost已經(jīng)降到了162，性能提升還是非常明顯。

小結(jié)

????通過(guò)以上說(shuō)明和測(cè)試，可以看到clustering factor也是索引健康的一個(gè)重要判斷的標(biāo)準(zhǔn)。其值越低越好。它會(huì)影響CBO選擇正確的執(zhí)行計(jì)劃。但是注意一點(diǎn)，clustering factor總是趨勢(shì)與不斷惡化的。

轉(zhuǎn)載于:https://blog.51cto.com/newcourage/1942656

總結(jié)

以上是生活随笔為你收集整理的Oracle:集群因子的全部?jī)?nèi)容，希望文章能夠幫你解決所遇到的問(wèn)題。

如果覺(jué)得生活随笔網(wǎng)站內(nèi)容還不錯(cuò)，歡迎將生活随笔推薦給好友。