Friends...
DB: Oracle11gR2
OS: Linux
I have created package with couple of procedure, procedure executes alter table move... , index rebuild command on database.
I'm doing below
Run ksh shell script -> execute procedure
Procedure runs alter table, rebuild index commands on database
Procedure completes
Shell script ends.
I can generate logfile for the shell script but whatever gets executed by procedure doesn't get recorded inside shell script logfile. I understood since db session created by procedure it won't record anything in shell logfile.
So how can I
Record everything in logfile which is executed by both procedures in the same package?
Also trying to put current datetime within procedure dmbs_out.put_line command?
Is it possible to run both procedure after connecting database once instead of 2 time connecting database and executing procedure?
There might be table/table partition move syntax error but I'm only trying to trap when table move started and when finished with datetime so to identify total time taken.
*** ksh script
#!/bin/ksh
...
...
...
$LOG_FILE = move_tbs.log
echo -e "set serveroutput on\n exec move_tbs.moveTable;"|$ORACLE_HOME/bin/sqlplus/#db_alias | head -l
echo -e "set serveroutput on\n exec move_tbs.moveTablePart;"|$ORACLE_HOME/bin/sqlplus/#db_alias | head -l
DB Package / Procedure
*** Procedure
create or replace package move_all
procedure moveTable
dbms_output.put_line("CURRENT TIME" 'alter table '|| owner || '.' || table_name || 'move');
Execute immediate 'alter table '|| owner || '.' || table_name || 'move';
dbms_output.put_line("COMPLETED TIME" : CURRENT_TIME);
end moveTable;
-------------------------------------------
procedure moveTablePart
dbms_output.put_line("CURRENT TIME" 'alter table '|| owner || '.' || table_name || 'move');
Execute immediate 'alter table '|| owner || '.' || table_name || 'move partition';
dbms_output.put_line("COMPLETED TIME" : CURRENT_TIME);
end moveTablePart;
end move_all
/
You could just put both exec commands in your echo construct. But you can use a 'heredoc' to simplify running both together, and it's a easier to read and maintain too. Something like:
LOG_TBS_MOVE=move_tbs.log
(
$ORACLE_HOME/bin/sqlplus -s -l user/passwd#db_alias <<!EOF
set serveroutput on
exec move_tbs.moveTable;
exec move_tbs.moveTablePart;
exit
!EOF
) > $LOG_TBS_MOVE
The herdoc start and end markers in this example !EOF - have to match exactly. They can be anything you like as long as there's no chance of anything inside the heredoc accidentally ending it. And the end marker has to be at the start of a line, it can't be indented.
The parentheses around the SQL*Plus and heredoc can enclose multiole commands and all output from within them goes into the log. They aren't really necessary here as there is only one command inside but it's a fairly clear way of doing the redirection, I think.
I'm only putting stdout into the $LOG_TBS_MOVE file; anything on stderr (which will not include any SQL errors) will still go to screen or to your main log if you redirect stderr for that.
To show the time in your output, don't enclose the current_time part in quotes, use string concatenation, and use the right function:
dbms_output.put_line(to_char(sysdate, 'HH24:MI:SS') ||
'alter table '|| owner || '.' || table_name || 'move');
dbms_output.put_line('COMPLETED TIME: ' || to_char(sysdate, 'HH24:MI:SS'));
Or you could display the time from the shell instead, within the parentheses; that would also then go into the same log file.
You haven't shown a username or password in your SQL*Plus calls; you probably just hid them, but if you are connecting as SYS via /, you really shouldn't be creating objects in that schema. Create a new schema and work in that.
Related
I am using the oracle version.
Oracle Database 11g Release 11.2.0.1.0
I accidentally ran the command o; in oracle delveloper.
The result is as below.
The PL/SQL procedure completed successfully.
not spooling currently
The sqlcl_int_runme alias has been removed.
I don't know what I did....
First of all, there seems to be no problem with basic table CRUD.
Has anyone had this experience?
I need an explanation of what happened...
It's an alias.
We copied over some popular commands from postgresql to SQLcl, one of those was 'o'
From the post docs
\o or \out [ filename ] \o or \out [ |command ] Arranges to save
future query results to the file filename or pipe future results to
the shell command command. If no argument is specified, the query
output is reset to the standard output.
If the argument begins with |, then the entire remainder of the line
is taken to be the command to execute, and neither variable
interpolation nor backquote expansion are performed in it. The rest of
the line is simply passed literally to the shell.
“Query results” includes all tables, command responses, and notices
obtained from the database server, as well as output of various
backslash commands that query the database (such as \d); but not error
messages.
SQL> alias
\! \? \c \cd \d \dp \dt \dt+ \e \echo \encoding \i
\o \p \prompt \q \qecho \r \save \timing \w \z clear cls
cpu fuzzy gglag locks sessions tables tables2 topsql
SQL> alias list \o
\o NULLDEFAULTS psql - desc \o [FILE_NAME] - turn spool log file on (or off if no FILE_NAME given)
--------------------------------------------------------------------------------------------------
Declare
maxpos number:=null;
BEGIN
if (:sqlcl_int_first is null) then
:sqlcl_int_runme:='spool off';
else
:sqlcl_int_runme:='spool '||:sqlcl_int_first||' ';
end if;
end;
/
alias NULLDEFAULTS sqlcl_int_runme=:sqlcl_int_runme;
sqlcl_int_runme
alias drop sqlcl_int_runme
To see it in action...
SQL> set sqlformat csv
SQL> o stackoverflow.csv
PL/SQL procedure successfully completed.
Alias sqlcl_int_runme dropped
SQL> select * from regions;
"REGION_ID","REGION_NAME"
1,"Europe"
2,"Americas"
3,"Asia"
4,"Middle East and Africa"
SQL> o
PL/SQL procedure successfully completed.
Alias sqlcl_int_runme dropped
SQL> !dir stackoverflow.csv
Volume in drive C is System
Volume Serial Number is F897-6A6F
Directory of c:\sqlcl\22.2.1\sqlcl\bin
08/30/2022 08:09 AM 170 stackoverflow.csv
1 File(s) 170 bytes
0 Dir(s) 190,156,173,312 bytes free
SQL> !type stackoverflow.csv
Alias sqlcl_int_runme dropped
"REGION_ID","REGION_NAME"
1,"Europe"
2,"Americas"
3,"Asia"
4,"Middle East and Africa"
PL/SQL procedure successfully completed.
I have created a script that creates Oracle users and grants them roles. I am unable to find a way to ignore the "user exists" error:
ORA-01920: user name '' conflicts with another user or role name.
I understand that when the script is ran, it is possible that the user already exists, but I want to ignore any returned errors. Is this possible?
My Oracle code:
CREATE USER "John" PROFILE "DEFAULT" IDENTIFIED BY "temppassword" ACCOUNT UNLOCK;
Edit:
This question is not asking how to create a user if it doesn't exist. This question is asking how to ignore "the user exists" error. According to a previously asked question, the top answer stated
In general, Oracle scripts simply execute the CREATE statement, and if
the object already exist, you'll get an error indicating that, which
you can ignore. This is whaat all the standard Oracle deployment
scripts do.
It isn't clear how you're running your script, but assuming its via SQL*Plus you can modify the behaviour when an error is encountered with the whenever sqlerror command.
If your script is setting that to exit at the moment, or you're picking that up from a startup script (login.sql, glogin.sql) you can change it back, or modify it temporarily:
...
-- do not stop on error
WHENEVER SQLERROR CONTINUE;
CREATE USER "John" PROFILE "DEFAULT" IDENTIFIED BY "temppassword" ACCOUNT UNLOCK;
-- to stop when later errors are encountered
WHENEVER SQLERROR EXIT FAILURE;
ALTER USER ...
You'll still see the ORA-01920 in the output but it will continue on to execute the next statement. This pattern is also useful for a protective drop of a schema object before attempting to create it.
Why can't you find if the user exists first?
SELECT COUNT(*)
INTO V_count
from ALL_USERS
where username = 'YourUserName'
IF v_count = 0 THEN
--create the user
--execute the grants
ELSE
---log that the user already exists
END IF;
SET SERVEROUTPUT ON;
DECLARE
TYPE t_list IS TABLE OF VARCHAR2 (30);
l_list t_list := t_list ('X0', 'X1', 'X2');
e_user_already_exists EXCEPTION;
PRAGMA EXCEPTION_INIT (e_user_already_exists, -1920);
BEGIN
FOR l_iterator IN 1 .. l_list.COUNT LOOP
DBMS_OUTPUT.PUT_LINE ('Creating user ' || l_list (l_iterator));
BEGIN
EXECUTE IMMEDIATE 'CREATE USER "' || l_list (l_iterator) || '" PROFILE DEFAULT IDENTIFIED BY "WELCOME" ACCOUNT UNLOCK';
EXECUTE IMMEDIATE 'GRANT SOME_APPLICATION_ROLE TO ' || l_list (l_iterator);
EXCEPTION
WHEN e_user_already_exists THEN
DBMS_OUTPUT.PUT_LINE ('User exists, ignored');
WHEN OTHERS THEN
RAISE;
END;
END LOOP;
END;
/
How to execute copy from command inside a plsql block?
E.g. I have copy from test/test#test insert emp using select * from emp;
How can I call this in a plsql block? I have tried with
execute immediate 'copy from test/test#test insert emp using select * from emp';
However when I execute my script which has plsql block gives me
ORA-00900: invalid SQL statement
How can I resolve this issue
COPY is a SQL*Plus command. So it only works in the SQL*Plus client. Find out more.
EXECUTE IMMEDIATE is a PL/SQL command to run dynamic calls, and it only recognises SQL and PL/SQL.
"I am executing sqlscript from sqlplus"
Yes, but you are calling COPY in an anonymous block, so that's with a PL/SQL scope; which means PL/SQL and SQL only.
The way to do this is with a shell script. These are operating system dependent, but something like this would work on a Linux environment.
#!/bin/bash
echo Please enter local Username:
read USERNAME
echo "Please enter local Password:"
read -s PASS
SID=${ORACLE_SID}
if [ "${ORACLE_SID}" != 'TEST' ]
then
sqlplus -s -l $USERNAME/$PASS#$SID << EOF
copy from test/test#test insert emp using select * from emp
exit
EOF
else
echo "Can't copy from TEST to TEST"
fi
Obviously this is just a wild guess at what your program actually does, but I hope you can understand the principle.
In a plsql code if we directly use the command as follows shall serve the similar output
begin
insert into emp1 select * from emp;
end;
emp1 is target table
emp is source table
There are similar ask where one wants to create blank structure or structure with data for backup kindly of activity.Refer link https://oracle-concepts-learning.blogspot.com/2019/09/copy-table-structure-or-data.html
1) Creating blank structure from existing table
--Execute on sql prompt
begin
execute immediate 'create table emp1 as select * from emp where 1=2';
end;
--Execute on sql prompt
select count(1) from emp1;
2) Creating structure from existing table with data
--Execute on sql prompt
begin
execute immediate 'create table emp1 as select * from emp';
end;
--Execute on sql prompt
select count(1) from emp1;
I have fairly simply code ..running in Oracle Virtualbox. However for some reason it is not displaying pl/sql output.
Here is code snippet
SQL> set serveroutput on
SQL> list
1 Create or Replace procedure mytz
2 IS
3 v_mytz TIMESTAMP WITH TIME ZONE DEFAULT '2013-05-05 12:00:00 AM';
4 BEGIN
5 DBMS_OUTPUT.PUT_LINE ('Default timestamp is ' );
6* end mytz ;
SQL> /
Procedure created.
SQL>
Is there anything I need to do special to see the output on SQL prompt ?
You have to actually run the procedure, not just create it, e.g.:
set serverputput on
exec mytz;
The set serveroutput SQL*Plus command has to be in the session the procedure is executed, not the one where it is created (if they are different).
You are not showing the value of your variable at the moment; maybe you wanted this?
dbms_output.put_line('Default timestamp is: ' || v_mytz);
Need a way to redirect PL/SQL program error message to a log file when executing it in sqlplus.
Say the PL/SQL program is named send_2012.sql and it has the following exception block
EXCEPTION
WHEN NO_DATA_FOUND
THEN
var_err := 'Data not found. ';
WHEN OTHERS
THEN
var_err := 'Error in '
|| $$plsql_unit
|| ' | '
|| SQLERRM
|| ' | '
|| 'Details: '
|| DBMS_UTILITY.format_error_backtrace;
END;
To run the PL/SQL program in a KornShell (ksh) script, I have:
sqlplus some_username/'some_password' #some_database \
#/some/directory/send_2012.sql \
$parameter1 $paramenter2
Suppose error occurs when executing send_2012.sql, how can I redirect the error message from var_err to /some/log/directory/log_send_2012.txt?
Much appreciated.
Setup your script like this:
-- test.sql script run from sqlplus
set serveroutput on
set echo on
WHENEVER SQLERROR EXIT SQL.SQLCODE
spool on
spool test.log
declare
l_val date;
begin
select sysdate into l_val from dual where 1=0;
exception
when others then raise;
end;
/
spool off
log into sqlplus from that directory and run:
SQL>#test.sql
You'll find the exceptions in the log file (test.log).
I worked around the logging issue, here is what I did:
Within the pl/sql program I inserted DBMS_PUTLINE("error messages goes here, etc"); to both the program body and exception sections.
When calling sqlplus from a Korn shell script, I used a regular output redirect to log pl/sql exceptions:
sqlplus some_username/'some_password' #some_database \
#/some/directory/send_2012.sql \
$parameter1 $paramenter2 \
> /some/log/directory/send_2012.log
What I did may not be the best solution. Wrap Spool before and after your pl/sql program may give you more options on formatting, but it may also include the output result (say from the select statement) when system executes the program successfully.