主流数据库 数据字典

AUTHOR: CenLiang

DATE: 2020/04/24

REV: 2.0

Oracle、MySQL、PostgreSQL、DB2、Hive 数据字典

1、Oracle 数据字典

## 1.1-oracle-表
SELECT b.username,
a.table_name,
a.tablespace_name,
a.num_rows,
a.status,
a.logging
dba_tables a,
(SELECT username, user_id, account_status, created
dba_users
WHERE account_status
username , SYSTEM, SYSMAN, DBSNMP, CTXSYSORDER account_status WHERE a.owner b.username
a.tablespace_name
a.num_rows
ORDER a.owner, a.tablespace_name, a.table_name;

## 1.2-oracle-分区表
SELECT b.username,
a.table_name,
a.tablespace_name,
a.num_rows,
a.partition_name,
a.partition_position
dba_tab_partitions a,
(SELECT username, user_id, account_status, created
dba_users
WHERE account_status
username , SYSTEM, SYSMAN, DBSNMP, CTXSYSORDER account_status WHERE a.table_owner b.username
ORDER b.username, a.table_name,a.partition_position;

## 1.3-oracle-字段
SELECT b.username,
a.table_name,
a.column_id,
a.column_name,
a.data_type,
a.data_length,
a.nullable
dba_tab_columns a,
(SELECT username, user_id, account_status, created
dba_users
WHERE account_status
username , SYSTEM, SYSMAN, DBSNMP, CTXSYSORDER account_status WHERE a.owner b.username
ORDER b.username, a.table_name, a.column_id;

2、MySQL 数据字典

## 2.1-mysql-表
SELECT
    TABLE_SCHEMA AS '库名',
    TABLE_NAME AS '表名',
    TABLE_COMMENT AS '表说明',
    TABLE_ROWS AS '数据量',
    CREATE_TIME AS '创建日期'
FROM
    information_schema.`TABLES`
WHERE
    TABLE_SCHEMA NOT IN (
        'mysql',
        'information_schema',
        'performance_schema',
                'sys'
    )
ORDER BY
    TABLE_SCHEMA,
    TABLE_NAME;

## 2.2-mysql-字段
SELECT
    TABLE_SCHEMA AS '库名',
    TABLE_NAME AS '表名',
    a.TABLE_COMMENT AS '表说明',
    b.ORDINAL_POSITION AS '列ID',
    b.COLUMN_NAME AS '字段名',
    b.COLUMN_TYPE AS '数据类型',
    b.IS_NULLABLE AS '允许为空',
    b.COLUMN_KEY AS '主键',
    b.COLUMN_DEFAULT AS '默认值',
    b.EXTRA AS '其他',
    b.COLUMN_COMMENT AS '字段说明'
FROM
    information_schema.`TABLES` a
JOIN information_schema.`COLUMNS` b USING (TABLE_SCHEMA,TABLE_NAME)
WHERE
    a.TABLE_SCHEMA NOT IN (
        'mysql',
        'information_schema',
        'performance_schema',
                'sys'
    )
GROUP BY
    TABLE_SCHEMA,
    TABLE_NAME,
    b.ORDINAL_POSITION
ORDER BY
    TABLE_SCHEMA,
    TABLE_NAME,
    b.ORDINAL_POSITION;

3、PostgreSQL 数据字典

## 3.1-pg-表
SELECT
    pt.tableowner AS "用户名",
    it.table_catalog AS "库名",
    it.table_schema AS "模式名",
    it."table_name" AS "表名",
--     nsp.oid as nspoid,
--     pc.oid as pcoid,
--     pd.objoid,
--     pd.objsubid,
    pd.description AS "表说明"
FROM
    information_schema.tables it
    LEFT JOIN pg_catalog.pg_tables pt ON ( pt.schemaname = it.table_schema AND pt.tablename = it."table_name" )
    LEFT JOIN pg_catalog.pg_namespace nsp ON (pt.schemaname = nsp.nspname)
    JOIN pg_catalog.pg_class pc ON ( nsp.oid = pc.relnamespace AND pt.tablename = pc.relname )
    LEFT JOIN pg_catalog.pg_description pd ON ( pc.oid = pd.objoid AND pd.objsubid = 0 )
WHERE
    it.table_schema NOT IN ( 'information_schema', 'pg_catalog', 'pg_toast' )
ORDER BY
    it.table_schema,
    it."table_name";

## 3.2-pg-字段
SELECT
    pt.tableowner AS "用户名",
    ic.table_catalog AS "库名",
--     nsp.oid AS nspoid,
    ic.table_schema AS "模式名",
--     pc.oid AS pcoid,
--     pa.attrelid AS paattrelid,
--     pd.objoid,
    ic."table_name" AS "表名",
    pd_tab.description AS "表说明" ,
--     pa.attnum,
--     pd.objsubid,
    ic.ordinal_position AS "列ID",
    ic."column_name" AS "字段名",
    ic.udt_name AS "数据类型",
    ic.is_nullable AS "允许为空",
    ic.column_default AS "默认值",
    pd_col.description AS "字段说明"
FROM
    information_schema."columns" ic
--     LEFT JOIN information_schema.tables it ON ( ic.table_schema = it.table_schema AND ic."table_name" = it."table_name" )
    LEFT JOIN pg_catalog.pg_tables pt ON ( ic.table_schema = pt.schemaname AND ic."table_name" = pt.tablename )
    LEFT JOIN pg_catalog.pg_namespace nsp ON ( ic.table_schema = nsp.nspname )
    JOIN pg_catalog.pg_class pc ON ( nsp.oid = pc.relnamespace AND ic."table_name" = pc.relname )
    LEFT JOIN pg_catalog.pg_description pd_tab ON ( pc.oid = pd_tab.objoid AND pd_tab.objsubid = 0 )
    LEFT JOIN pg_catalog.pg_attribute pa ON ( pc.oid = pa.attrelid AND ic.ordinal_position = pa.attnum )
    LEFT JOIN pg_catalog.pg_description pd_col ON ( pa.attrelid = pd_col.objoid AND pa.attnum = pd_col.objsubid AND pd_col.objsubid > 0 )
WHERE
    ic.table_schema NOT IN ( 'information_schema', 'pg_catalog', 'pg_toast' )
ORDER BY
    ic.table_schema,
    ic."table_name",
    ic.ordinal_position;

4、DB2 数据字典

## 4.1-db2-模式
SELECT
    (
        SELECT DISTINCT
            (table_catalog)
        FROM
            sysibm.tables WITH ur) AS TABLE_CATALOG,
    SCHEMATA.SCHEMANAME            AS schema_name,
    SCHEMATA.CREATE_TIME
FROM
    syscat.SCHEMATA
WHERE
    SCHEMATA.SCHEMANAME NOT IN ('SYSIBM',
                                'SYSCAT',
                                'SYSFUN',
                                'SYSSTAT',
                                'SYSPROC',
                                'SYSIBMADM',
                                'SYSIBMINTERNAL',
                                'SQLJ',
                                                            'SYSTOOLS',
                                                            'NULLID',
                                                            'SYSIBMTS',
                                'SYSPUBLIC');

## 4.2-db2-表
SELECT
    it.TABLE_CATALOG,
    ct.TABSCHEMA AS schema_name,
    ct.TABNAME   AS table_name,
    ct.REMARKS   AS table_comment,
    ct.CREATE_TIME,
    ct.LASTUSED,
    ct.COLCOUNT AS column_num,
    ct.CARD     AS table_card
FROM
    syscat.TABLES ct
JOIN
    SYSIBM.TABLES it
ON
    ct.TABSCHEMA=it.TABLE_SCHEMA
AND ct.TABNAME=it.TABLE_NAME
WHERE
    ct.TABSCHEMA NOT IN ('SYSIBM',
                         'SYSCAT',
                         'SYSFUN',
                         'SYSSTAT',
                         'SYSPROC',
                         'SYSIBMADM',
                                                'SQLJ',
                                                'SYSTOOLS',
                                                'NULLID',
                         'SYSIBMINTERNAL',
                         'SYSIBMTS',
                         'SYSPUBLIC')
ORDER BY
    ct.TABSCHEMA,
    ct.TABNAME;

## 4.3-db2-字段
SELECT
    ic.TABLE_CATALOG,
    cc.TABSCHEMA AS schema_name,
    cc.TABNAME   AS table_name,
    ct.REMARKS   AS table_comment,
    ic.ORDINAL_POSITION,
    cc.COLNAME  AS column_name,
    cc.TYPENAME AS data_type,
    cc.LENGTH,
    ic.IS_NULLABLE,
    cc.DEFAULT,
    cc.REMARKS AS column_comment,
    cc.HIGH2KEY,
    cc.LOW2KEY,
    cc.COLCARD AS column_card
FROM
    syscat.TABLES ct,
    syscat.COLUMNS cc,
    SYSIBM.COLUMNS ic
WHERE
    ct.TABSCHEMA=cc.TABSCHEMA
AND ct.TABNAME=cc.TABNAME
AND cc.TABSCHEMA=ic.TABLE_SCHEMA
AND cc.TABNAME=ic.TABLE_NAME
AND cc.COLNAME =ic.COLUMN_NAME
AND cc.TABSCHEMA NOT IN ('SYSIBM',
                         'SYSCAT',
                         'SYSFUN',
                         'SYSSTAT',
                         'SYSPROC',
                         'SYSIBMADM',
                                               'SQLJ',
                                               'SYSTOOLS',
                                               'NULLID',
                         'SYSIBMINTERNAL',
                         'SYSIBMTS',
                         'SYSPUBLIC')
ORDER BY
    cc.TABSCHEMA,
    cc.TABNAME,
    ic.ORDINAL_POSITION;

5、Hive 数据字典

## 5.1-hive-表
SELECT
    d.OWNER_NAME AS '用户名',
    -- DB_ID,
    d.`NAME` AS '库名',
    -- t.TBL_ID,
    -- t.SD_ID,
    t.TBL_NAME AS '表名',
    t1.`comment` AS '表说明',
    t1.numRows AS '数据量',
    FROM_UNIXTIME(t.CREATE_TIME) AS '创建时间'
FROM
    hive.DBS d JOIN hive.TBLS t USING(DB_ID)
JOIN (SELECT
TBL_ID,
MAX(CASE tp.PARAM_KEY WHEN 'comment' THEN tp.PARAM_VALUE ELSE NULL END) AS `comment`,
MAX(CASE tp.PARAM_KEY WHEN 'numRows' THEN tp.PARAM_VALUE ELSE NULL END) AS `numRows`
FROM hive.TABLE_PARAMS tp
GROUP BY TBL_ID ) t1 USING(TBL_ID)
ORDER BY d.`NAME`,t.TBL_NAME;

## 5.2-hive-字段
SELECT
    d.OWNER_NAME AS '用户名',
    -- DB_ID,
    d.`NAME` AS '库名',
    -- t.TBL_ID,
    -- t.SD_ID,
    t.TBL_NAME AS '表名',
    t1.`comment` AS '表说明',
    c1.INTEGER_IDX AS '列ID',
    c1.COLUMN_NAME AS '字段名',
    c1.TYPE_NAME AS '数据类型',
    c1.`COMMENT` AS '字段说明'
FROM
    hive.DBS d JOIN hive.TBLS t USING(DB_ID)
JOIN (SELECT
TBL_ID,
MAX(CASE tp.PARAM_KEY WHEN 'comment' THEN tp.PARAM_VALUE ELSE NULL END) AS `comment`
FROM hive.TABLE_PARAMS tp
GROUP BY TBL_ID ) t1 USING(TBL_ID)
JOIN (SELECT
    s.SD_ID,
    CD_ID,
    c.INTEGER_IDX,
    c.COLUMN_NAME,
    c.TYPE_NAME,
    c.`COMMENT`
FROM
    hive.SDS s
JOIN hive.COLUMNS_V2 c USING (CD_ID)) c1 USING(SD_ID)
ORDER BY d.`NAME`,t.TBL_NAME,c1.INTEGER_IDX;