分区表工具脚本

1 prompt <p>当前用户下，表大小超过10个GB未建分区的
2 select segment_name,
3        segment_type,
4        sum(bytes) / 1024 / 1024 / 1024 object_size
5   from user_segments
6 WHERE segment_type = ‘TABLE‘
7 group by  segment_name, segment_type
8 having sum(bytes) / 1024 / 1024 / 1024 >= 10
9 order by object_size desc;

 1 prompt <p>查询当前用户下，失效-普通索引 
 2 select t.index_name,
 3        t.table_name,
 4        blevel,
 5        t.num_rows,
 6        t.leaf_blocks,
 7        t.distinct_keys
 8   from user_indexes t
 9 where status = ‘INVALID‘;
10 
11 prompt <p>查询当前用户下的失效-分区索引
12 select t1.blevel,
13        t1.leaf_blocks,
14        t1.INDEX_NAME,
15        t2.table_name,
16        t1.PARTITION_NAME,
17        t1.STATUS
18   from user_ind_partitions t1, user_indexes t2
19 where t1.index_name = t2.index_name
20    and t1.STATUS = ‘UNUSABLE‘;

 1 --以下脚本可以分析分区表记录不平衡的情况，同时也可以从侧面发现由于疏于管理，大量当前数据进入默认分区的情况
 2 /*
 3 
 4    注：这个语句不一定准确，尤其是在表未及时收集统计信息的时候
 5    
 6 */
 7 
 8 --统计信息系统一般会自动收集，这只是首次建成表后需要操作一下，以方便测试
 9 exec dbms_stats.gather_table_stats(ownname => ‘LJB‘,tabname => ‘RANGE_PART_TAB‘,estimate_percent => 10,method_opt=> ‘for all indexed columns‘,cascade=>TRUE) ;  
10 
11 
12 --当前用户下，某个分区的记录数是平均记录数的2倍以上
13 set linesize 266
14 col table_name format a20
15 select table_name,
16        max(num_rows),
17        trunc(avg(num_rows),0),
18        sum(num_rows),
19        case when sum(num_rows),0 then 0,else trunc(max(num_rows) / sum(num_rows),2) end,
20        count(*)
21   from user_tab_partitions
22  group by table_name
23 having max(num_rows) / sum(num_rows) > 2 / count(*);
24 
25 --也可用来作为判断查询当前用户下有因为疏于分区管理导致无大量数据进了建默认分区的参考。
26 
27 select table_name,
28        partition_name,
29        num_rows  
30  from user_tab_partitions
31  where table_name = ‘RANGE_PART_TAB‘
32  order by num_rows desc;
33 
34 
35 --请试验<分区类型_范围分.sql>后，体会上述脚本

 1 prompt <p>当前用户下分区最多的前10个对象
 2 select *
 3   from (select  table_name, count(*) cnt
 4           from user_tab_partitions
 5          group by  table_name
 6          order by cnt desc)
 7 where rownum <= 10;
 8 
 9 prompt <p>当前用户下分区个数超过100个的表
10 select  table_name, count(*) cnt
11   from user_tab_partitions
12  having count(*) >= 100
13 group by table_name, table_name
14 order by cnt desc;
15 
16 --或者如下更方便
17 select table_name, partitioning_type, subpartitioning_type
18   from user_part_tables
19  where partition_count > 100;

1 select table_name, 
2        partitioning_type, 
3        subpartitioning_type, 
4        partition_count
5   from user_part_tables
6  where subpartitioning_type <> ‘NONE‘;
7  
8  
9 select count(*) from user_part_tables where  subpartitioning_type <> ‘NONE‘;

 1 select table_name, 
 2        column_name, 
 3        data_type
 4   from user_tab_columns
 5  where data_type in ( ‘LONG‘,‘CHAR‘);
 6 
 7 
 8 
 9 
10 
11 
12 
13 
14 
15 
16 
17 ------------------------------------------------------------------------------------
18 
19 drop table t_long purge;
20 create table t_long (id int, name long);
21 drop table t_char purge;
22 create table t_char (id int, address char(10));
23 set linesize 266
24 col table_name  format a25
25 col column_name format a25
26 col data_type   format a20
27 
28 SQL> select table_name,
29   2         column_name,
30   3         data_type
31   4    from user_tab_columns
32   5   where data_type in ( ‘LONG‘,‘CHAR‘);
33 
34 TABLE_NAME                COLUMN_NAME             DATA_TYPE
35 ------------------------- ----------------------------------
36 T_CHAR                    ADDRESS                   CHAR
37 T_LONG                    NAME                      LONG

 1 --查看当前数据库哪些对象外键没建索引
 2 select table_name,
 3        constraint_name,
 4        cname1 || nvl2(cname2, ‘,‘ || cname2, null) ||
 5        nvl2(cname3, ‘,‘ || cname3, null) ||
 6        nvl2(cname4, ‘,‘ || cname4, null) ||
 7        nvl2(cname5, ‘,‘ || cname5, null) ||
 8        nvl2(cname6, ‘,‘ || cname6, null) ||
 9        nvl2(cname7, ‘,‘ || cname7, null) ||
10        nvl2(cname8, ‘,‘ || cname8, null) columns
11   from (select b.table_name,
12                b.constraint_name,
13                max(decode(position, 1, column_name, null)) cname1,
14                max(decode(position, 2, column_name, null)) cname2,
15                max(decode(position, 3, column_name, null)) cname3,
16                max(decode(position, 4, column_name, null)) cname4,
17                max(decode(position, 5, column_name, null)) cname5,
18                max(decode(position, 6, column_name, null)) cname6,
19                max(decode(position, 7, column_name, null)) cname7,
20                max(decode(position, 8, column_name, null)) cname8,
21                count(*) col_cnt
22           from (select substr(table_name, 1, 30) table_name,
23                        substr(constraint_name, 1, 30) constraint_name,
24                        substr(column_name, 1, 30) column_name,
25                        position
26                   from user_cons_columns) a,
27                user_constraints b
28          where a.constraint_name = b.constraint_name
29            and b.constraint_type = ‘R‘
30          group by b.table_name, b.constraint_name) cons
31  where col_cnt > ALL
32  (select count(*)
33           from user_ind_columns i
34          where i.table_name = cons.table_name
35            and i.column_name in (cname1, cname2, cname3, cname4, cname5,
36                 cname6, cname7, cname8)
37            and i.column_position <= cons.col_cnt
38          group by i.index_name)
39 
40 
41 查询所有含外键的表
42 
43 select count(*),TABLE_NAME,c_constraint_name from (
44 select a.table_name,       
45        substr(a.constraint_name, 1, 30) c_constraint_name,
46        substr(a.column_name, 1, 30) column_name,
47        position,
48        b.owner,
49        b.constraint_name,
50        b.constraint_type
51   from user_cons_columns a, user_constraints b
52  where a.constraint_name = b.constraint_name
53    and b.constraint_type = ‘R‘ )
54    group by TABLE_NAME,c_constraint_name
55 
56