Oracle自动生成根据源表迁移到HIVE创表语句
SQL脚本
============================================
自动生成表脚本文件
============================================
select 'CREATE TABLE IF NOT EXISTS &OWNER.' || table_name || chr(13) || '(' ||
RTRIM(XMLAGG(XMLPARSE(CONTENT(column_name || chr(9) || DATA_TYPE || chr(10) || ',') WELLFORMED) order by column_id)
.GETCLOBVAL(),
',') || ')' || 'COMMENT ' || '"' || tab_comments || '"' ||
chr(10) || 'PARTITIONED BY (col type)' || chr(10) ||
'CLUSTERED BY (col) SORTED BY(col) INTO BUCKETS' || chr(10) ||
'STORED AS parquet' || chr(10) ||
'tblproperties(''parquet.compression''=''snappy'');' as sq,
table_name
from (select col.table_name,
col.COLUMN_NAME,
decode(col.DATA_TYPE,
'NUMBER',
'DECIMAL',
'TIMESTAMP(6)',
'TIMESTAMP') || CASE
WHEN DATA_TYPE = 'NUMBER' THEN
'(' || nvl(to_char(DATA_PRECISION), 22) || ',' ||
to_char(NVL(COL.data_scale, 0)) || ')'
WHEN LOWER(col.COLUMN_NAME) in ('inserttime', 'updatetime') then
' TIMESTAMP'
when data_type IN ('DATE', 'TIMESTAMP(6)') THEN
' TIMESTAMP'
ELSE
' STRING'
END || ' COMMENT ' || '''' || cms.COMMENTS || '''' || CHR(13) AS DATA_TYPE,
col.COLUMN_ID,
tms.COMMENTS as tab_comments
from all_tab_columns col
join all_col_comments cms
on col.TABLE_NAME = cms.table_name
and col.COLUMN_NAME = cms.COLUMN_NAME
AND col.OWNER = cms.OWNER
join all_tab_comments tms
on col.TABLE_NAME = tms.TABLE_NAME
AND col.OWNER = tms.OWNER
where COL.TABLE_NAME like upper('%&tab_name%')
AND regexp_substr(tms.table_name, '[0-9]+') is null
and col.OWNER = '&src_owner')
group by table_name, tab_comments
ORDER BY table_name; 测试过程
执行脚本SQL-F8
填写参数
OWNER：表示迁移到的数据库名称
tab_name：要迁移的表名
src_owner：表示表原来所在的数据库名称查看结果
测试结果
=========================================================
创建表语句
=========================================================
CREATE TABLE IF NOT EXISTS DEFAULT.SSA_TP_TDIVIDENDDETAIL_HIS
(D_CDATE TIMESTAMP COMMENT ''
,C_CSERIALNO STRING COMMENT ''
,D_REGDATE TIMESTAMP COMMENT ''
,D_DATE TIMESTAMP COMMENT ''
,C_FUNDACCO STRING COMMENT ''
,C_TRADEACCO STRING COMMENT ''
,C_FUNDCODE STRING COMMENT ''
,C_SHARETYPE STRING COMMENT ''
,C_AGENCYNO STRING COMMENT ''
,C_NETNO STRING COMMENT ''
,F_TOTALSHARE DECIMAL(16,2) COMMENT ''
,F_UNITPROFIT DECIMAL(7,4) COMMENT ''
,F_TOTALPROFIT DECIMAL(16,2) COMMENT ''
,F_TAX DECIMAL(16,2) COMMENT ''
,C_FLAG STRING COMMENT ''
,F_REALBALANCE DECIMAL(16,2) COMMENT ''
,F_REINVESTBALANCE DECIMAL(16,2) COMMENT ''
,F_REALSHARES DECIMAL(16,2) COMMENT ''
,F_FARE DECIMAL(16,2) COMMENT ''
,D_LASTDATE TIMESTAMP COMMENT ''
,F_NETVALUE DECIMAL(7,4) COMMENT ''
,F_FROZENBALANCE DECIMAL(16,2) COMMENT ''
,F_FROZENSHARES DECIMAL(16,2) COMMENT ''
,F_INCOMETAX DECIMAL(9,4) COMMENT ''
,C_RESERVE STRING COMMENT ''
,D_REQUESTDATE TIMESTAMP COMMENT ''
,C_SHARECLASS STRING COMMENT ''
,C_SENDTYPE STRING COMMENT ''
)COMMENT ""
PARTITIONED BY (col type)
CLUSTERED BY (col) SORTED BY(col) INTO BUCKETS
STORED AS parquet
tblproperties('parquet.compression'='snappy'); 根据自己的情况自行选择是分区还是分桶，按照哪些字段。
注意分区的话，要将分区字段注释。