• Status: Solved
  • Priority: Medium
  • Security: Public
  • Views: 94
  • Last Modified:

Best way to rerun DDL in oracle database within shell script on linux

Hi,
we have a set of DDL which I execute every time as part of a process on oracle database.

1. First time when we execute, it creates and alter tables.
2. On all the subsequent executions it will throw an error which I ignore safely. Since they already exist in the database.
3. I agree this is not the best way to re-execute. I should have some kind of check before re-executing each ddl.
4. Recently it is taking time between 20 mins to 50 mins to even throw error depends on the environments

Pls help what is the best way to check if column/index already exist and skip this.

thanks in advance

Below are the sample script with only few ddl and their error msg and elapsed time.

alter table SAMPLE.EIM_GROUP
  add "REL_SEQ_FLG" VARCHAR2(1 CHAR); 

  ERROR at line 2:
ORA-01430: column being added already exists in table


Elapsed: 00:00:00.01

alter table SAMPLE.EIM_GROUP
  add "SCORE" NUMBER(10,0);
ERROR at line 2:
ORA-01430: column being added already exists in table


Elapsed: 00:00:00.01

CREATE INDEX SAMPLE.CHILD_X1_X ON SAMPLE.CON_CHILD (ATTRIB_07,UCM_TYPE_CD,OBJECT_TYPE,OBJECT_TABLE,CREATED) TABLESPACE USER_TBLSPACE;
ERROR at line 1:
ORA-01652: unable to extend temp segment by 128 in tablespace TEMP


Elapsed: 00:07:21.57

Open in new window

0
enthuguy
Asked:
enthuguy
1 Solution
 
Abhimanyu SuriSr Database EngineerCommented:
Column exists error is not taking time as per the given example, it is create index.

Error is for TEMP tablespace being full, please increase the size of it.

You can check from user_tab_columns/dba_tab_columns/all_tab_columns using a combination of schema,table_name,column_name if the column being added already exists or not and take action accordingly.

For example :

SELECT column_name
    FROM dba_tab_columns
   WHERE owner = 'SCHEMA_NAME'
     AND table_name = 'TABLE_NAME'
     AND column_name = 'COLUMN_NAME' --> Name of column being added

Even, oracle itself does the same check along with many others with exception handling.
Thats how it knows that column already exist

Excerpt from 10046 trace :

select count(FA#) from SYS_FBA_TRACKEDTABLES where OBJ# = 512779 and bitand(FLAGS, 128)=0
....
....
 execute immediate m_stmt into m_o_stmt;
  if (m_owner!='MDSYS' and m_owner!='SYS' and m_type='TABLE' and m_event='ALTER')
  then
    m_stmt:='select column_name from dba_tab_columns where owner=:1 and table_name=:2';
    open m_cur for m_stmt using m_owner,m_name;
    loop
      fetch m_cur into m_column;
      exit when m_cur%NOTFOUND;
      m_stmt:='select sdo_geor_def.isDropColumn(:1) from dual';
1
 
Mark GeerlingsDatabase AdministratorCommented:
The best way to handle this is likely in a PL\SQL procedure (or a group of procedures)  That would allow you to execute a query first, then depending on whether the query determines the object already exists or not, the PL\SQL procedure can decide to execute the DDL command or not.  

Some of the data dictionary views that would be the most helpful to write your queries to select from are:
dba_tables
dba_tab_columns
dba_indexes

If you don't have access to these "dba_..." views, use the corresponding "all_..." views instead.

And, be aware that in PL\SQL you can't execute DDL commands directly like you can in a *.sql script.  In PL\SQL, these can only be executed inside the "execute immediate" command or in procedures from the supplied DBMS_SQL package.
0

Featured Post

Free Tool: SSL Checker

Scans your site and returns information about your SSL implementation and certificate. Helpful for debugging and validating your SSL configuration.

One of a set of tools we are providing to everyone as a way of saying thank you for being a part of the community.

Tackle projects and never again get stuck behind a technical roadblock.
Join Now