Greenplum 常用数据字典

一、数据库集群信息

1、gp_segment_configration

2、pg_filespace_entry

这两个表是在pg_global表空间下面的，是全局表。

用来查看集群segment信息，比如segment个数。

二、常用数据字典表

1、pg_class

保存了所有表、视图、序列、索引元数据信息，每个DDL/DML操作都必须跟这个表发生关系。

2、pg_attribute

记录字段的内容

3、gp_distribution_policy

记录表的分布键

4、pg_statistic 和 pg_stats

数据库中表的统计信息保存在pg_statistic中

pg_stats可以方便帮我们查看pg_statistic的内容

5、pg_partition

记录分区表的信息

6、pg_partition_rule

分区表的分区规则

7、pg_partitions

三、字典表的应用

1、查看表的信息

1）从catalog中查询

2）从information_schema查询

select   *  from information_schema.tables where table_name ='';

1、获取字段信息

1）从catalog中查询

SELECT a.attname,pg_catalog.format_type(a.atttypid,( pg_catalog.pg_attribute a, a.atttypmod) AS data_type
FROM pg_catalog.pg_attribute a,
 (SELECT c. oid 
 FROM pg_catalog.pg_class c 
 LEFT JOIN pg_catalog.pg_namespace n 
 ON n.oid = c.relnamespace 
 WHERE c.relname = 'pg_class' 
 AND n.nspname = 'pg_catalog' 
 ) b  
WHERE a.attrelid = b.oid 
AND a.attnum > 0 
AND NOT a.attisdropped ORDER BY a.attnum; 

使用regclass就会简化很多：
SELECT a.attname,pg_catalog.format_type(a.atttypid, a.atttypmod) AS data_type
FROM pg_catalog.pg_attribute a
WHERE a.attrelid ='pg_catalog.pg_class'::regclass
AND a.attnum > 0
AND NOT a.attisdropped ORDER BY a.attnum;

2）从information_schema查询

select * from information_schema.columns where table_name = '';

2、获取表的分布键

select a.attrnums[i.i],b.attname,a.localoid::regclass

from gp_distribution_policy a,

(select generate_series(1,10))i (i),

pg_attribute b

where a.attrnums[i.i) is not null

and a.localoid=b.attrelid

and a.attrnums[i.i]=b.attnum

and a.localoid='public.cxfa2 '::regclass

order by i.i;

3、获取一个视图定义

testDB=# create table cxfa( a int) distributed by (a);
CREATE TABLE
testDB=# create view v_cxfa as select * from cxfa;
CREATE VIEW
testDB=# select pg_get_viewdef('v_cxfa', true);
SELECT cxfa.a FROM cxfa;
(1 row)

4、查询备注信息

1）获取表的备注

select COALESCE(description, * *) as comment from pg_description where objoid=‘cxfa*::regclass and objsubid=0;

2）获取字段的备注

select  b.attname as columnname, COALESCE(a.description,'') as comment

from      pg_catalog.pg_description a,pg_catalog.pg_attribute b

where  objoid='cxfa'::regclass

and a.objoid=b.attrelid

and a.objsubid=b.attnum;

5、获取数据据库建表语句

1）调用plpythonu

CREATE PROCEDURAL LANGUAGE plpythonu;

2）创建存储过程

create or replace function get_table_structure(tablename text)    # 表名格式为 tableschema.table_name
    returns text
as $$
    try:
        table_name = tablename.lower().split('.')[1]

        talbe_schema=tablename.lower().split('.')[0]

    except (IndexError):

        return 'Please in put "tableschema.table_name"'

    get_table_oid="select oid,reloptions,relkind from pg_class where oid='%s'::regclass"%(tablename)

    try:

        rv_oid=plpy.execute(get_table_oid,5)      #plpy.execute(query [, max-rows])

        if not rv_oid:

            return 'Did not find any relation named"'+tablename +'".'

    except (Error):

        return 'Did not find any relation named"'+tablename +'".'

    table_oid=rv_oid[0]['oid']

    rv_reloptions=rv_oid[0]['reloptions'] # 访问方法特定的选项，使用"keyword=value"格式的字符串

    rv_relkind=rv_oid[0]['relkind']

    create_sql="";

    table_kind='table';

    if rv_relkind !='r' and rv_relkind !='v':  # r:普通表， v:视图， c:复合类型， t:表， o:内部 AO 节点文件，

        plpy.error('%s is not table or view'%(tablename));

    elif rv_relkind=='v':

        get_view_def="select pg_get_viewdef(%s,'t') as viewdef;" % (table_oid)

        rv_viewdef=plpy.execute(get_view_def);

        create_sql='create view %s as \n' % (tablename)

        create_sql += rv_viewdef[0]['viewdef']+'\n';

        table_kind='view'

    else:

       #获取行

      get_columns="select a.attname,pg_catalog.format_type(a.atttypid,a.atttypmod),\

       (select substring(pg_catalog.pg_get_expr(d.adbin,d.adrelid) for 128) \

        from pg_catalog.pg_attrdef d where d.adrelid=a.attrelid and d.adnum=a.attnum and a.atthasdef) \

        as default,a.attnotnull as isnull from pg_catalog.pg_attribute \

        a where a.attrelid= %s and a.attnum >0 and not a.attisdropped order by a.attnum;" % (table_oid);

        rv_columns=plpy.execute(get_columns)

       

        #获取分布键

        get_table_distribution1="select attrnums from pg_catalog.gp_distribution_policy t where localoid = '" + table_oid + "' "

        rv_distribution1=plpy.execute(get_table_distribution1,500)

        rv_distribution2=''

        if rv_distribution1 and rv_distribution1[0]['attrnums']:

            get_table_distribution2="select attname from pg_attribute where attrelid='"+table_oid+"' and attnum in (" + str(rv_distribution1[0]['attrnums']).strip('{').strip('}').strip('[').strip(']')+")"

            rv_distribution2=plpy.execute(get_table_distribution2,500)

 

        create_sql='create table %s (\n' % (tablename)

        #获取索引

        get_index="select pg_get_indexdef(indexrelid) as indexdef from pg_index where indrelid=%s" % (table_oid);

        rv_index=plpy.execute(get_index);

 

        get_parinfo1="select attname as columnname from pg_attribute where attnum =(select paratts[0] from pg_partition where parrelid=%s) and attrelid=%s;"%(table_oid,table_oid);

        get_parinfo2=""" select pp.parrelid,prl.parchildrelid,case when pp.parkind='h'::"char" then 'hash'::text when pp.parkind='r'::"char" then 'range'::text when pp.parkind='l'::"char" then 'list'::text else null::text end as partitiontype,pg_get_partition_rule_def(prl.oid,true) as partitionboundary from pg_partition pp,pg_partition_rule prl where pp.paristemplate=false and pp.parrelid = %s and prl.paroid = pp.oid order by prl.parname; """ % (table_oid)

        v_par_parent=plpy.execute(get_parinfo1);

        v_par_info=plpy.execute(get_parinfo2);

        max_column_len=10

        max_type_len=4

        max_modifiers_len=4

        max_default_len=4

        for i in rv_columns:

            if i['attname']:

                if max_column_len < i['attname'].__len__():

                    max_column_len=i['attname'].__len__()

            if i['format_type']:

                if max_type_len < i['format_type'].__len__():

                    max_type_len=i['format_type'].__len__()

            if i['default']:

                if max_type_len < i['default'].__len__():

                    max_default_len=i['default'].__len__()

        first=True

        for i in rv_columns:

            if first==True:

                split_char=' ';

                first=False

            else:

                split_char=',';

            if i['attname']:

                create_sql += " " + split_char + i['attname'].ljust(max_column_len+6)+''

            else:

                create_sql += "" + split_char + ' '.ljust(max_column_len+6)

            if i['format_type']:

                create_sql += ' ' + i['format_type'].ljust(max_type_len +2)

            else:

                create_sql += ' ' + ' '.ljust(max_type_len+2)

            if i['isnull'] and i['isnull']:

                create_sql += ' ' + ' not null '.ljust(8)

            if i['default']:

                create_sql += ' default ' + i['default'].ljust(max_default_len+6)

            create_sql += "\n"

        create_sql += ")"

 

        if rv_reloptions:

            create_sql +=" with ("+str(rv_reloptions).strip('{').strip('}').strip('[').strip(']') +")\n"

            create_sql = create_sql.replace("'",'')

        if rv_distribution2:

            create_sql += 'Distributed by ('

            for i in rv_distribution2:

                create_sql += i['attname'] + ','

            create_sql =create_sql.strip(',')+')'

        elif rv_distribution1:

            create_sql += 'Distributed randomly\n'

        if v_par_parent:

            partitiontype=v_par_info[0]['partitiontype'];

            create_sql +='\nPARTITION BY '+ partitiontype + "("+v_par_parent[0]['columnname']+")\n(\n";

            for i in v_par_info:

                create_sql +=" " +i['partitionboundary']+',\n';

            create_sql=create_sql.strip(',\n');

            create_sql+="\n)"

        create_sql+=";\n\n"

        for i in rv_index:

            create_sql += i['indexdef']+';\n'

 

        get_table_comment="select 'comment on %s %s is '''|| COALESCE (description,'')|| '''' as comment from pg_description where objoid=%s and objsubid=0;" % (table_kind,tablename,table_oid)

        get_column_comment="select 'comment on column %s.'||b.attname ||' is ''' || COALESCE(a.description,'')|| ''' ' as comment from pg_catalog.pg_description a,pg_catalog.pg_attribute b where objoid=%s and a.objoid=b.attrelid and a.objsubid=b.attnum;" % (tablename,table_oid)

        rv_table_comment=plpy.execute(get_table_comment);

        rv_column_comment=plpy.execute(get_column_comment);

 

        for i in rv_table_comment:

            create_sql += i['comment']+';\n'

        for i in rv_column_comment:

            create_sql +=i['comment']+';\n'

        return create_sql;

 

$$ LANGUAGE plpythonu;

3）使用例子

testdb=# SELECT get_table_structure('public.tb1_partition_range_yyyymmdd');

                                                get_table_structure                                               

-------------------------------------------------------------------------------------------------------------------

 create table public.tb1_partition_range_yyyymmdd (                                                               

   id               numeric，

   yyyymmdd         date                                                                                          

 ) with (appendonly=true, compresslevel=5)                                                                        

 Distributed by (id)                                                                                              

 PARTITION BY range(yyyymmdd)                                                                                     

 (                                                                                                                

  PARTITION p20120811 START ('2012-08-11'::date) END ('2012-08-12'::date) WITH (appendonly=true, compresslevel=5),

  PARTITION p20120812 START ('2012-08-12'::date) END ('2012-08-13'::date) WITH (appendonly=true, compresslevel=5) 

 );                                                                                                               

                                                                                                                  

 CREATE INDEX idx_yyyymmdd ON tb1_partition_range_yyyymmdd USING btree (yyyymmdd);    

 

5、查看表的依赖关系

pg_depend