Making a script that generates an output file dynamic

Question

I've created the following script to run it from within SQL*Plus on my PC to generate a report regarding specific details about all my databases in one file. The idea is to get 1 result of all my databases.
Spool C:\output.txt;
-- start DB1
CONNECT system/pass@DB1;
COLUMN host_name format A18;
select * from v$database;
set line 380;
set pagesize 50;
select * from v$session order by username;
DISCONNECT;
-- end DB1
-- start DB2
CONNECT system/pass@DB2;
.......................same code as for DB1..................................
DISCONNECT;
-- end DB2
-- start DB3
CONNECT system/pass@DB3;
.......................same code as for DB1..................................
DISCONNECT;
-- end DB3
Spool off;
The above script, for instance, generates the name of the database and the sessions details of all my databases.
I have 40 DBs and they all have the same password for user system.
My question, is there a way to re-write this script to achieve the same thing but in a dynamic way. ie: the database connection string becomes dynamic, so that I write the code only once and if the code changes, I change it only one time instead of 40 times.
Notes: I'd like to run the script from SQL*Plus on my Window PC, I know that the same thing can be achieved by running a script on the DB server by using /etc/oratab, but actually these 40 DBs are on 10 different DB servers and there is no oratab used.
Many thanks for any tips,
Thomas

Answer 1

By awk on unix, or gawk or mawk on Windows (download http://www.klabaster.com/freeware.htm or ftp://garbo.uwasa.fi/pc/unix/gawk2156.zip) it's a easy task.
File select.sql (your query which you want run)
spool E:\Scripts\Sql\&1..log
select table_name from user_tables;
spool off
exitFile db.txt :
#col1 = ORACLE_HOME, col2=user, col3=pwd, col4=dbname, col5=script to run (cool if you run different scritps)
e:\oracle\ora92 scott tiger DEMO92 E:\Scripts\Sql\select.sql
e:\oracle\ora102 scott demo102 DEMO102 E:\Scripts\Sql\select.sqlFile select.awk (program calling by main prog)
{ print "set ORACLE_HOME="$1 }
{print "set ORACLE_SID=" $4}
{print "cd %ORACLE_HOME%"}
{print "cd bin"}
{print "sqlplus -s " $2"/"$3" @"$5" "$4}File run_select.cmd (the program which you will run)
gawk\gawk -f select.awk db.txt>run_select.cmd
run_select.cmdWell, now you have pfile and prog, you can work by simple enter name run_awk.cmd or scheduling by task windows :
E:\Scripts\Sql>run_awk.cmd    <-- you have finish your job, all lines after this one are automatic
E:\Scripts\Sql>gawk\gawk -f select.awk db.txt 1>run_select.cmd
E:\Scripts\Sql>run_select.cmd
E:\Scripts\Sql>set ORACLE_HOME=e:\oracle\ora92
E:\Scripts\Sql>set ORACLE_SID=DEMO92
E:\Scripts\Sql>cd e:\oracle\ora92
E:\oracle\ora92>cd bin
E:\oracle\ora92\bin>sqlplus -s scott/tiger @E:\Scripts\Sql\select.sql DEMO92
TABLE_NAME
A
BONUS
C
DEPT
EMP
HARSIMRAT
PLAN_TABLE
SALGRADE
T
T2
TAB1
11 rows selected.
E:\oracle\ora92\bin>set ORACLE_HOME=e:\oracle\ora102
E:\oracle\ora92\bin>set ORACLE_SID=DEMO102
E:\oracle\ora92\bin>cd e:\oracle\ora102
E:\oracle\ora102>cd bin
E:\oracle\ora102\BIN>sqlplus -s scott/demo102 @E:\Scripts\Sql\select.sql DEMO102
TABLE_NAME
DEPT
EMP
BONUS
SALGRADE
C
TAB1
TBL
MATRIX
TAB2
TAB3
TAB4
TABLE_NAME
TAB5
TEMP
TBL7
TBL6
IZZA
TBLA
TBLB
TBLC
TMP_LOOKUPS
TMP_HIER
TBL8
TABLE_NAME
TBL9
MV1
EMP2
THE_TABLE
26 rows selected.
E:\oracle\ora102\BIN>Two logfiles were generate DEMO92.log and DEMO102.log.
Now you can declinate this procedure to the infinity.
Nicolas.
Message was edited by:
N. Gasparotto
You have an alternative with mawk which seems more fast, command are same, just replace gawk by mawk.
Message was edited by:
N. Gasparotto