Selective excel table import/export

Question

I need to run an SQL script that will selectively (select * from <TABLE> where <"TransactionRunDate" older than 6 months from SystemDate>) remove all rows from three table and simultaneously write them into excel sheets. Then, I need an SQL script that can restore the data from these Excel tables into the Oracle tables (on demand)
How can I do this without right clicking import/export on the GUI (SQLDeveloper), but with SQL script alone?
Edited by: user10403078 on Oct 7, 2008 11:32 PM

Answer 1

I would half agree with daniel and half not agree.
You can't but you can. As for whether you should is a different matter.
I say you can because it's possible to do something like you are looking for. You could even get it to write out Excel workbooks with worksheets using Microsofts office XML format and then use an ODBC connection to read the data back again into an Oracle database using Heterogeneous services, but it's complicated and not easily demonstrated with what little time I have.
However, an alternative is to write the data out to CSV files and then use External tables to read those CSV files back again. Example...
Firstly, here's my table with some data in it, some of which is more than 6 months old...
SQL> select * from table1;
        ID CREATED_DATE        TXT
         1 13/12/2007 09:31:08 Old Fred
         2 01/02/2008 09:31:08 Old Bob
         3 22/03/2008 09:31:08 Old Jim
         4 11/05/2008 09:31:08 Young Fred
         5 30/06/2008 09:31:08 Young Bob
         6 19/08/2008 09:31:08 Young Jim
         7 08/10/2008 09:31:08 New Me
7 rows selected.Now, I create a DELETE trigger on that table so if any rows are deleted this will fire. To allow for rows less than 6 months old to be deleted without being archived to file I put a check in the trigger too...
SQL> ed
Wrote file afiedt.buf
  1  create or replace trigger trg_del
  2  after delete on table1
  3  for each row
  4  declare
  5    v_dir  VARCHAR2(30) := 'TEST_DIR';
  6    v_file VARCHAR2(30) := 'table1_archive.csv';
  7    v_fh   UTL_FILE.FILE_TYPE;
  8  begin
  9    IF months_between(sysdate, :OLD.created_date) >= 6 THEN
10      -- Only archive a deleted row if it's greater than 6 months old
11      v_fh := UTL_FILE.FOPEN(v_dir, v_file, 'a', 32767);
12      UTL_FILE.PUT_LINE(v_fh, TO_CHAR(:OLD.ID,'fm9999')||','||TO_CHAR(:OLD.CREATED_DATE,'YYYYMMDDHH24MISS')||',"'||:OLD.TXT||'"');
13      UTL_FILE.FCLOSE(v_fh);
14    END IF;
15* end;
SQL> /
Trigger created.So now, I have a trigger that will write deleted rows of data out in CSV format to a file called table1_archive.csv.
Now I can create an external table that can read that CSV file...
SQL> ed
Wrote file afiedt.buf
  1  CREATE TABLE table1_archive (
  2         id            NUMBER,
  3         created_date  DATE,
  4         txt           VARCHAR(200)
  5         )
  6  ORGANIZATION EXTERNAL (
  7    TYPE oracle_loader
  8    DEFAULT DIRECTORY TEST_DIR
  9      ACCESS PARAMETERS (
10      RECORDS DELIMITED BY NEWLINE
11      BADFILE 'bad_%a_%p.bad'
12      LOGFILE 'log_%a_%p.log'
13      FIELDS TERMINATED BY ',' OPTIONALLY ENCLOSED BY '"'
14      MISSING FIELD VALUES ARE NULL
15      REJECT ROWS WITH ALL NULL FIELDS
16        (id
17        ,created_date date "yyyymmddhh24miss"
18        ,txt)
19      )
20      LOCATION ('table1_archive.csv')
21    )
22  PARALLEL
23  REJECT LIMIT 0
24* NOMONITORING
SQL> /
Table created.
{code}
And now we can prove it works...
Firstly, delete rows from the table that are 6 months or more old...
{code}
SQL> delete from table1
  2  where months_between(sysdate, created_date) >= 6;
3 rows deleted.
SQL> select * from table1;
        ID CREATED_DATE        TXT
         4 11/05/2008 09:31:08 Young Fred
         5 30/06/2008 09:31:08 Young Bob
         6 19/08/2008 09:31:08 Young Jim
         7 08/10/2008 09:31:08 New Me
{code}
So those rows have been deleted, now let's check they're in the archive CSV file...
{code}
SQL> select * from table1_archive;
        ID CREATED_DATE        TXT
         1 13/12/2007 09:31:08 Old Fred
         2 01/02/2008 09:31:08 Old Bob
         3 22/03/2008 09:31:08 Old Jim
SQL>
{code}
There they are, and that CSV file can also be read by Excel if required.
Let's just check that we don't archive rows that are deleted if they are less than 6 months old...
{code}
SQL> delete from table1 where id = 6;
1 row deleted.
SQL> select * from table1;
        ID CREATED_DATE        TXT
         4 11/05/2008 09:31:08 Young Fred
         5 30/06/2008 09:31:08 Young Bob
         7 08/10/2008 09:31:08 New Me
SQL> select * from table1_archive;
        ID CREATED_DATE        TXT
         1 13/12/2007 09:31:08 Old Fred
         2 01/02/2008 09:31:08 Old Bob
         3 22/03/2008 09:31:08 Old Jim
SQL>
{code}
So that row was deleted without being archived.  Just what I intended.