當(dāng)前位置:首頁 >  站長 >  數(shù)據(jù)庫 >  正文

postgresql 刪除重復(fù)數(shù)據(jù)的幾種方法小結(jié)

 2021-04-25 17:12  來源: 腳本之家   我來投稿 撤稿糾錯

  域名預(yù)訂/競價,好“米”不錯過

在使用PG數(shù)據(jù)庫的這段時間,總結(jié)了三種刪除重復(fù)數(shù)據(jù)的方法,其中最容易想到的就是最常規(guī)的刪除方法,但此方法性能較差,刪數(shù)據(jù)耗時較久,雖容易實(shí)現(xiàn),但性能太差,影響寫數(shù)據(jù)的速率。

另外就是被使用的group by刪除方法,效率較高。

還有一種是剛發(fā)現(xiàn)的,還沒有驗(yàn)證,現(xiàn)在就總結(jié)下這三種刪除方法,并驗(yàn)證各自的執(zhí)行效率。

首先創(chuàng)建一張基礎(chǔ)表,并插入一定量的重復(fù)數(shù)據(jù)。

  test=# create table deltest(id int, name varchar(255));
  CREATE TABLE
  test=# create table deltest_bk (like deltest);
  CREATE TABLE
  test=# insert into deltest select generate_series(1, 10000), 'ZhangSan';
  INSERT 0 10000
  test=# insert into deltest select generate_series(1, 10000), 'ZhangSan';
  INSERT 0 10000
  test=# insert into deltest_bk select * from deltest;

 

常規(guī)刪除方法

最容易想到的方法就是判斷數(shù)據(jù)是否重復(fù),對于重復(fù)的數(shù)據(jù)只保留ctid最?。ɑ蜃畲螅┑哪菞l數(shù)據(jù),刪除其他的數(shù)據(jù)。

test=# explain analyse delete from deltest a where a.ctid <> (select min(t.ctid) from deltest t where a.id=t.id);
                               QUERY PLAN
  -----------------------------------------------------------------------------------------------------------------------------
  Delete on deltest a (cost=0.00..195616.30 rows=1518 width=6) (actual time=67758.866..67758.866 rows=0 loops=1)
    -> Seq Scan on deltest a (cost=0.00..195616.30 rows=1518 width=6) (actual time=32896.517..67663.228 rows=10000 loops=1)
     Filter: (ctid <> (SubPlan 1))
     Rows Removed by Filter: 10000
     SubPlan 1
      -> Aggregate (cost=128.10..128.10 rows=1 width=6) (actual time=3.374..3.374 rows=1 loops=20000)
         -> Seq Scan on deltest t (cost=0.00..128.07 rows=8 width=6) (actual time=0.831..3.344 rows=2 loops=20000)
            Filter: (a.id = id)
            Rows Removed by Filter: 19998
  Total runtime: 67758.931 ms
  test=# select count(*) from deltest;
  count
  -------
  10000
  (1 行記錄)

 

可以看到,id相同的數(shù)據(jù),保留ctid最小的那條,其他的刪除。相當(dāng)于把deltest表中的數(shù)據(jù)刪掉一半,耗時達(dá)到67s多。相當(dāng)慢。

group by刪除方法

第二種方法為group by方法,通過分組找到ctid最小的數(shù)據(jù),然后刪除其他數(shù)據(jù)。

  test=# truncate table deltest;
  TRUNCATE TABLE
  test=# insert into deltest select * from deltest_bk;
  INSERT 0 20000
  test=# explain analyse delete from deltest a where a.ctid not in (select min(ctid) from deltest group by id);
                               QUERY PLAN
  ----------------------------------------------------------------------------------------------------------------------------------
  Delete on deltest a (cost=131.89..2930.46 rows=763 width=6) (actual time=30942.496..30942.496 rows=0 loops=1)
    -> Seq Scan on deltest a (cost=131.89..2930.46 rows=763 width=6) (actual time=10186.296..30814.366 rows=10000 loops=1)
     Filter: (NOT (SubPlan 1))
     Rows Removed by Filter: 10000
     SubPlan 1
      -> Materialize (cost=131.89..134.89 rows=200 width=10) (actual time=0.001..0.471 rows=7500 loops=20000)
         -> HashAggregate (cost=131.89..133.89 rows=200 width=10) (actual time=10.568..13.584 rows=10000 loops=1)
            -> Seq Scan on deltest (cost=0.00..124.26 rows=1526 width=10) (actual time=0.006..3.829 rows=20000 loops=1)
   Total runtime: 30942.819 ms
  (9 行記錄)
  test=# select count(*) from deltest;
   count
  -------
  10000
  (1 行記錄)

 

可以看到同樣是刪除一半的數(shù)據(jù),使用group by的方式,時間節(jié)省了一半。但仍含需要30s,下面試一下第三種刪除操作。

新的刪除方法

在postgres修煉之道這本書中,作者提到一種效率較高的刪除方法, 在這里驗(yàn)證一下,具體如下:

  test=# truncate table deltest;
  TRUNCATE TABLE
  test=# insert into deltest select * from deltest_bk;
  INSERT 0 20000                            
  test=# explain analyze delete from deltest a where a.ctid = any(array (select ctid from (select row_number() over (partition by id), ctid from deltest) t where t.row_number > 1));
                               QUERY PLAN
  ----------------------------------------------------------------------------------------------------------------------------------
  Delete on deltest a (cost=250.74..270.84 rows=10 width=6) (actual time=98.363..98.363 rows=0 loops=1)
  InitPlan 1 (returns $0)
   -> Subquery Scan on t (cost=204.95..250.73 rows=509 width=6) (actual time=29.446..47.867 rows=10000 loops=1)
      Filter: (t.row_number > 1)
      Rows Removed by Filter: 10000
      -> WindowAgg (cost=204.95..231.66 rows=1526 width=10) (actual time=29.436..44.790 rows=20000 loops=1)
         -> Sort (cost=204.95..208.77 rows=1526 width=10) (actual time=12.466..13.754 rows=20000 loops=1)
            Sort Key: deltest.id
            Sort Method: quicksort Memory: 1294kB
            -> Seq Scan on deltest (cost=0.00..124.26 rows=1526 width=10) (actual time=0.021..5.110 rows=20000 loops=1)
  -> Tid Scan on deltest a (cost=0.01..20.11 rows=10 width=6) (actual time=82.983..88.751 rows=10000 loops=1)
     TID Cond: (ctid = ANY ($0))
  Total runtime: 98.912 ms
  (13 行記錄)
  test=# select count(*) from deltest;
  count
  -------
  10000
  (1 行記錄)

 

看到上述結(jié)果,真讓我吃驚了一把,這么快的刪除方法還是首次看到,自己真實(shí)孤陋寡聞,在這里要膜拜一下修煉之道這本書的大神作者了。

補(bǔ)充:pgsql 刪除表中重復(fù)數(shù)據(jù)保留其中的一條

1.在表中(表名:table 主鍵:id)增加一個字段rownum,類型為serial

2.執(zhí)行語句:

delete from table where rownum not in(
select max(rownum) from table group by id
)

 

3.最后刪除rownum

以上為個人經(jīng)驗(yàn),希望能給大家一個參考,也希望大家多多支持。如有錯誤或未考慮完全的地方,望不吝賜教。

文章來源:腳本之家

來源地址:https://www.jb51.net/article/205415.htm

申請創(chuàng)業(yè)報(bào)道,分享創(chuàng)業(yè)好點(diǎn)子。點(diǎn)擊此處,共同探討創(chuàng)業(yè)新機(jī)遇!

相關(guān)文章

熱門排行

信息推薦