SQL Plus: retrieve the commands

11g R2 (11.2.0.3) on Linux. At the SQL Plus prompt such as SQL>, I typed some commands, say, select from test;* Sometime I want to re-use the query, so I need to retype in select from test.* Is there a way to retrieve the earlier issued commands back?
Also, I can not use the BACK arrow key. Suppose I want to insert or delete a character in the middle of the commandline, how can I move the cursor over the character?
Thanks

955300 wrote:
11g R2 (11.2.0.3) on Linux. At the SQL Plus prompt such as SQL>, I typed some commands, say, select from test;* Sometime I want to re-use the query, so I need to retype in select from test.* Is there a way to retrieve the earlier issued commands back?Yes. The utility is called rlwrap. You use it to run sqlplus. I alias it. So in my +.bash_profile+, I will have the following line:
alias sqlplus="rlwrap $ORACLE_HOME/sqlplus"It provides full command line recall (inside SQL*Plus) and command line editing.
There is no rlwrap rpm available for installation AFAIK. I'm downloading the source code and doing a manual build and install for all the server installations I do (I do not use Oracle Linux or RHEL as a desktop o/s - think for my Ubuntu desktop it installed via an <i>apt-get</i>). For a manual source code build, you need to install a couple of development rpm's as I recall, such as the readline lib.

Similar Messages

Does SQL Developer have any way to run the SQL*PLUS COPY FROM command?

set arraysize 100;
set copycommit 100;
COPY FROM SCOTT/TIGER@OtherDBinstance -
CREATE COMPANY -
USING SELECT * from COMPANY;

Thanks. I have been using the create table ABC as select * from ABC@otherdb but this often runs into temp segment issues for very large tables. The SQL*PLUS version allows me to have the copy commit save after each xxxxx lines and this seems to be able to handle any size table. Can copy commit be used from SQL Developer?

SQL*Plus Worksheet DEFINE command

**I hope this is the correct place to put this, I don't see a forum for SQL*Plus Worksheet.**
I am having some trouble with SQL*Plus Worksheet. The following call to a package works fine,
EXECUTE package_name.procedure_name ('string_name');where it is required that my argument passed to the procedure needs the quotes as part of the string. However, when I try to use a variable instead of hardcoding the 'string_name' value, ie
DEFINE string_name = 'string_name';
EXECUTE package_name.procedure_name (&string_name);I get an error message saying that my variable needs to be defined (the call was made passing the string w/o the quotes). How can I pass the entire string, quotes and all, to the package call? I've tried concatenating them to the variable in the DEFINE and EXECUTE commands but I can't seem to come up with the correct syntax.
Does anybody know how to do this?

Here is the current SQL*Plus Worksheet command, and output (works):
EXECUTE amanda.role_executor.drop_role ('read_only');
PL/SQL procedure successfully completed.So, now I try to pass the argument as a variable, instead of hardcoding the value to the call:
DEFINE role_name = 'read_only';
EXECUTE amanda.role_executor.drop_role (&read_only);
BEGIN amanda.role_executor.drop_role (read_only); END;
ERROR at line 1:
ORA-06550: line 1, column 39:
PLS-00201: identifier 'READ_ONLY' must be declared
ORA-06550: line 1, column 7:
PL/SQL: Statement ignoredAs you can see when I try to pass the string as a variable the single quotes get lost. How can I concatenate the single quotes to the string? I've tried everything I can think of but apparently Worksheet doesn't like the || symbols.

SQL*PLUS using START command - please help me to understand

Hello dear friends,
I met one issue that i can't understand. Could you please explain me what is wrong in my logic.
I have created sql file called DUMMY.sql
the file context is:
SELECT dummy
FROM dummy_tbl
WHERE dummy_col=&dummy_var;
I go to SQL*Plus DB version (10.2.0.4.0).
I promt command: START DUMMY 1
and expecting *&dummy_var* to be replaced with *1* and query be executed as it is described in Oracle SQL*PLUS reference ([http://download-west.oracle.com/docs/cd/B12037_01/server.101/b12170/ch6.htm#sthref1058|http://download-west.oracle.com/docs/cd/B12037_01/server.101/b12170/ch6.htm#sthref1058] ).
But instead i'm getting Enter value for dummy_var:
please help me to understand
thanks in advance.

You need to use &1 instead of &dummy_var.
The number references the position of the parameter to use;&1 is the first parameter, &2 would be a second parameter and so on.
SELECT dummy
FROM dummy_tbl
WHERE dummy_col=&1;

Redirect the output of SQL*Plus to the file

Hi all,
I have the following command to run:
sqlplus username/password @myfile.sql
I don't know how to redirect the output to the file.
Also in the script myfile.sql I have the defined variable, &VP and
How can I substitute this variable(&VP) at the run time.
Thanks,
JP

Assumed you like to replace your defined variable VP by number 4. SInce it is posintional parameter,instead of wrinting &VP, write
VP:=&1
sqlplus>spool redirect.txt
sqlplus>@myfile.sql 4
sqlplus>spool off

Having problems trying to run a sql script using the command prompt.

I place my script at the root c:
open run SQL Command Line.
I connected to the database using connet userid/password
Then I tried to execute the script by doing
@test_add_comments
I am getting the error message.
SQL> @test_add_comments
SP2-0310: unable to open file "test_add_comments.sql"
Am I suppose to place this sql somewhere else?
test_add_comments.sql is
COMMENT ON COLUMN myTable.ID IS 'Primary Key' ;
COMMENT ON COLUMN myTable.REVIEWLOG_ID IS 'REFERENCE TO SAFE REVIEW LOG TABLE' ;
COMMENT ON COLUMN myTable.FAC_ID IS 'REFERENCE TO ALL FACILITIES' ;
COMMENT ON COLUMN myTable.HALLBUILD_ID IS 'REFERENCE TO HALL BUILDING' ;

I figure it out.
open a sql worksheet and enter the falling commands in the run script.
@c:\your_script.sql
Execute it by clicking run script.

SQL to retrieve the records in subset

Dear Friends,
I have requirement to fetch 5 Million records from Oracle db and display using JAVA and JSP. Each record is 1 KB size.
The database table is online table used to store the transaction details, the rows will keep increasing. I am not allowed to use stored procedures also.
1) Fetching 5 Million record and displaying in 1 attempt is not feasible as JDBC has limitation and also the User has to wait too long to see the data.
I am thinking of processing the records in subset, for example.. Display first 10000 records then if the users wants next 10000 records fetch and display the next 10000 record.
Are there any other best practices to implement the above requirement.
I am not sure how to write the above SQL query? Can anyone please help me write the above SQL query?
I dont this I can use romnum as the rows are increasing.
Below is my Table structure and SQL query which returns 5 Million record
CREATE TABLE MESSAGE_LOG
SOURCE VARCHAR2(25 BYTE),
TARGET VARCHAR2(25 BYTE),
ENGINE VARCHAR2(100 BYTE),
PROCESS VARCHAR2(400 BYTE),
SERVICE VARCHAR2(100 BYTE),
LOGTIMESTAMP TIMESTAMP(6)
SELECT * FROM MESSAGE_LOG WHERE LOGTIMESTAMP BETWEEN TO_DATE('20-MAY-2010 01:00','DD-MON-YYYY HH24:MI') AND TO_DATE('20-MAY-2010 23:00','DD-MON-YYYY HH24:MI')
Thanks in advance,
Bob

user13109986 wrote:
I have requirement to fetch 5 Million records from Oracle db and display using JAVA and JSP. Why no one will look at five million results. And even if they wanted to it would take longer than the average human lifetime.
I am not allowed to use stored procedures also.Why, is there a business requirement that says the application should be as difficult to develop as possible, be unreliable and hard to maintain when things go wrong?
>
1) Fetching 5 Million record and displaying in 1 attempt is not feasible as JDBC has limitation and also the User has to wait too long to see the data.As the user will never read five million results, it doesn't matter
Try this link and come back when you have got to the 5,000,000 result of the 6,730,000 that are estimated to be there
http://www.google.com/#hl=en&q=oracle+procedure
Or you could list every thread in the forum and multiply the time it takes you to get to the last by 50.
I am thinking of processing the records in subset, for example.. Display first 10000 records then if the users wants next 10000 records fetch and display the next 10000 record.Try looking at 10,000 threads on this forum or 10,000 results from Google and see how long that takes.
Are there any other best practices to implement the above requirement.Yes, don't return more information than anyone could possibly find useful

SQL to retrieve the specific backuppiece to restore the archivelog/datafile

Hi,
Does anybody have any scripts handy to figure out which specific backuppiece to restore if I need to restore a specific archivelog/datafile? i.e.
1. I am using a controlfile and no rman catalog.
2. There is a gap in the Standby site and I need to scp multiple backup pieces to DR site to restore the archivelogs and apply them.
3. I want to query the primary database and find which specific backuppieces I need to scp to restore these files.
Is there a function/Procedure/Package/Table/View/SQL Query available to figure this out?
Thanks a lot.

Thanks Fahd for the reply.
But my question is a bit different. I know the archilogs that are needed and they have already been backed up by RMAN and deleted in primary. I need to restore them in DR but need to scp only those pieces which are holding the gap logs. I dont want to scp all of the backuppieces, just the ones holding multiple logs that I need to apply.
Hope it helps. Thanks for your effort though... Appreciated :)

How to compare two rows in PL/SQL and retrieve the values?

Hello,
I have two tables which have identical schemas, one table (tbl_store) is used to hold the latest version, and the other table (tbl_store_audit) holds previous versions. When the latest record is updated, it is inserted into the tbl_store_audit table as a revision, and the updated details are used as the latest record.
For example: The latest version is held in tbl_store, however the tbl_store_audit may hold 5 records which are the past records used before changes were made - these are seen as revisions.
I want to be able to compare what has changed between each revision in the tbl_store_audit table. For example: Out of the 10 columns, the change between revision 1 and revision 2 was the size from XL to XXL. The change between revision 3 and revision 4 was the size XS to M and price 4.99 to 10.99, and so on.
Eventually i will create an APEX report that will show the user the revision number and what was changed from and to.
I seen in a previous post i need to note my oracle version: Oracle version 10.2.0.4.0

Hi,
Like suggested already you should give some sample data and output.
Maybe you would like to have something like this:
-- Sample data
-- Note PK is the primairy key of the source table and rev are the revisions
with tbl_store_audit as
select 1 pk, 1 rev , 1 price , 'XXL' unit_size from dual union all
select 1 pk, 2 rev , 1 price,   'XL' unit_size from dual union all
select 1 pk, 3 rev , 1.4 price, 'XXL' unit_size from dual union all
select 2 pk, 1 rev , 1.4 price, 'XL' unit_size from dual union all
select 2 pk, 2 rev , 1.4 price, 'XL' unit_size from dual union all
select 2 pk, 3 rev , 1.4 price, 'XL' unit_size from dual union all
select 1 pk, 4 rev , 1 price , 'XL' unit_size from dual union all
select 1 pk, 5 rev , 1 price , 'XL' unit_size from dual union all
select 3 pk, 1 rev , 1.2 price, 'XL' unit_size from dual union all
select 3 pk, 2 rev , 1.2 price, 'XXL' unit_size from dual union all
select 4 pk, 1 rev , 1 price , 'XL' unit_size from dual
-- end of sample data
,tbl_store_audit_tmp as
select
pk
,rev
,'PRICE'          field_name
,to_char(price)   field_value
,to_char(lag(price,1) over (partition by pk order by rev) ) old_field_value
from
tbl_store_audit
union all
select
pk
,rev
,'UNIT_SIZE'           field_name
,to_char(UNIT_SIZE)    field_value
,to_char(lag(UNIT_SIZE,1) over (partition by pk order by rev) ) old_field_value
from
tbl_store_audit
-- include all other fields from the table here with it's own union all select ...
select
from
tbl_store_audit_tmp
where
field_value != old_field_value
PK REV FIELD_NAME FIELD_VALUE                              OLD_FIELD_VALUE
1   3 PRICE      1.4                                      1
1   4 PRICE      1                                        1.4
1   2 UNIT_SIZE XL                                       XXL
1   3 UNIT_SIZE XXL                                      XL
1   4 UNIT_SIZE XL                                       XXL
3   2 UNIT_SIZE XXL                                      XL
6 rows selected If you realy want to keep track of all the changes I think it would be better if you make a "after update trigger" on the base table that checks what changed and put that directly in the uadit table.
Regards,
Peter
Edited by: Peter vd Zwan on Aug 16, 2012 8:25 AM

SQL works from SQL*Plus command line but not as a DBMS_JOB submitted job

Oracle 10g 10.2
Got a procedure which does not run correctly as an Oracle job but runs fine as a SQL script.
There are no Oracle errors (or any errors) of any kind when the job does not run fine – it just does not update any rows. But when run as a SQL script – the same way it is run as a job – then the rows are updated.
Any ideas?

Good stuff....
See my replies to some of the questions in italics
a) Different NLS settings => The job uses the NLS settings of the session that created it. If you create it with some tool like TOAD, you might have a different environment than with sql*plus. Runs good in TOAD and SQL*Plus using the command-line feature - just 'acts funky' when submitted.
b) Interval issue. Are you sure the job was running? Isn't still running? The job runs successfully - even logs a successful message in our logging table.
c) User/priviledge issue. Sometimes a job needs direct grants whereus a procedure can be called with priviledges granted through roles. Don't know - need to check this out.
d) Transaction handling / Error Handling. The job runs into some error, but the error is supressed, because of bad exception handling. What is the value in the BROKEN column, when you do: select * from user_jobs; I would think that, since the errors are logged into a side table, an error would be found there. However, no errors are found. The BROKEN column is 'N'.

About the command 'HOST'

My environment: Oracle Workgroup Server for Windows NT (Version
7.3.4)
Can anyone, who can connect the Oracle database with a legal db
account through the SQL*Plus, run the command ‘HOST’ normally?
How to limit the usage of command ‘HOST’?
Thanks!

You can prevent the HOST command being used by inserting an
appropriate record into the product_user_profile table.
See the PUP section of your manual:
http://download-west.oracle.com/otndoc/oracle9i/901_doc/server.90
1/a88827/ape2.htm#634109
You may also find the command line option -RESTRICT (introduced
in 8.1.6) useful:
http://download-west.oracle.com/otndoc/oracle9i/901_doc/server.90
1/a88827/ch72.htm#1005809
- CJ

Very new to sql plus

I am using sql plus by the command line, I am using the ed mode. My problem is, I have two inserts one right after another. in note pad:
insert into table values
insert into table values
('d')
The problem is only one insert gets done.
What is the correct way to write this ?
Eric

If it is the same table, you could do something like:
SQL> create table mytable(col1 varchar2(10));
Table created.
SQL> insert into mytable(col1) values(&value);
Enter value for value: ' '
old   1: insert into mytable(col1) values(&value)
new   1: insert into mytable(col1) values(' ')
1 row created.
SQL> r
1* insert into mytable(col1) values(&value)
Enter value for value: 'd'
old   1: insert into mytable(col1) values(&value)
new   1: insert into mytable(col1) values('d')
1 row created.
SQL> select * from mytable;
COL1
d
SQL> set feedback on
SQL> /
COL1
d
2 rows selected.

Within SQL*Plus, get error code of host command

Hello everyone,
I am currently writing an SQL*Plus (Oracle 10g) that has a big logic and somewhere in the middle, I have to call a host command, which is an C++ function.
I call it using
SQL> HOST cd
SQL> HOST cd bins
SQL> HOST ./my_procedureBut, I need to get the return code of my procedure to see if it finished correctly?
I checked [the manual|http://download.oracle.com/docs/cd/B19306_01/server.102/b14357/ch4.htm#sthref883] but they don't say anything about return codes.
When I try:
SQL> host echo $?
0It always returns 0, even if I SIGTERM or SIGKILL the child process. Anyone has an idea on how is it possible to retrieve the return code?
Thanks

user13117585 wrote:
Hello guys and thanks for your answers.
Let me tell you a little more about my problem... I have a 3 steps process. The first step is done in the database. It updates a few tables. Then, when that step successfully ends, I need to start a second step. That second step is done on the server where I have my SQL Plus session. Because I actually have 2 different machines. The first one is where the database is running and the second one where I have SQL Plus (Oracle client 10g) and the program that I have to start in this second step. Then, once this process is finished and his return code is 0, then I have to execute the third step in the database.
Basically, I wanted to have something like:
VARIABLE returnCode
BEGIN
:returnCode := some_package.do_step_1;
-- do whatever I want with returnCode.
EXCEPTION WHEN ...
END;
HOST execute_external_program
BEGIN
IF external_job_succeed THEN
some_package.do_step_3
END IF;
EXCEPTION...
END;
/I wanted to use the Oracle Scheduler to create an EXTERNAL JOB. The problem is that the external program is not deployed on the same machine as the database.I know that on 11g, we have external remote jobs. But I'm on 10g and I have no agent on the server where the external program is. So, what are my options? SQL*Plus and the host command... Or I can also write a bash script to do that... Or any hight level programming language...
Any other idea?
Thanks again guys,move functionality of second step into PL/SQL procedure resident inside DB.

How to change the default directory from within SQL*Plus ?

Hello,
I want to change the default directory directly from within SQL*Plus to be able to launch my command files with simple instructions like :
@my_command.sql
If I haven't launched SQL*Plus from the directory containing the file my_command.sql, how do I change the default directory to the directory of my command files ?
I've searched through the Oracle documentation, there's no SET DEFAULTDIR or something like that. When I do a SHOW ALL, there's no variable containing the default directory.
If you have an idea...
regards,
Jérôme.

Hi JÃ©rÃ´me (Jé²´me ?),
You cannot directly do that, but you can use @@ instead of @ to run subscripts in the same directory as the superscript
# head /tmp/tl30/xxx/[xyz].sql
==> /tmp/tl30/xxx/x.sql <==
prompt call y
@y
prompt call z
@@z
quit
==> /tmp/tl30/xxx/y.sql <==
prompt i am in y
==> /tmp/tl30/xxx/z.sql <==
prompt i am in z
# pwd
/root
# sqlplus -s scott/tiger@lsc62 @/tmp/tl30/xxx/x.sql
call y
SP2-0310: Datei "y.sql" konnte nicht geÃ¶ffnet werden
call z
i am in zKind regards
Laurent Schneider
OCM DBA

Maximum length of SQL*Plus Command?

I'm typing a long query into SQL*Plus (Linux, version 9.2.0.1... yes, I know it's ancient).
If I run this query, all works fine:
select 'update myauditing_table set test_details='''||grantee||''' where myaudit_refno=''A.FE'';' from (select case when upper(value) in ('DBO','TRUE') then (select agg_concat(grantee) from dba_tab_privs where table_name='AUD$')
else (select 'AUDIT_TRAIL is set to: '||value from dual) end as grantee from v$parameter where upper(name)='AUDIT_TRAIL' order by grantee);
'UPDATEMYAUDITING_TABLESETTEST_DETAILS='''||GRANTEE||'''WHEREMYAUDIT_REFNO=''A.F
update myauditing_table set test_details='AUDIT_TRAIL is set to: DB' where myaud
it_refno='A.FE';(In other words, is SQL that generates other SQL: if AUDIT_TRAIL is set to something interesting, tell me who has access to the AUD$ table. And it's working fine... the agg_concat function mentioned is simply Tom Kyte's stragg function under another name, as lifted word-for-word from http://www.sqlsnippets.com/en/topic-11591.html)
Now this is what happens if I take out the linebreak after the mention of AUD$, so that the command is submitted as one single line of code:
SQL> select 'update myauditing_table set test_details='''||grantee||''' where myaudit_refno=''A.FE'';' from (select case when upper(value) in ('DBO','TRUE') then (select agg_concat(grantee) from dba_tab_privs where table_name='AUD$') else (select 'AUDIT_TRAIL is set to: '||value from dual) end as grantee from v$parameter where upper(name)='AUDIT_TRAIL' order by grantee);
SP2-0734: unknown command beginning "s set to: ..." - rest of line ignored.The error message indicates that the text has been 'split' at the 'AUDIT_TRAIL is set to...' bit, so that "s set to" is being treated as a new command, which is of course syntactically invalid. There is definitely no other change in text between the two versions, apart from the removal of a carriage return before the "else" statement.
Possibly a coincidence, but the first "s" in "s set to" appears at position 258 in the entire text... close to a possible 255 or 256 character limit, perhaps?
The same problem happens whether I run the command as a script (@mysql.sql) or typed in directly into a client SQL*Plus session. Is there some inherent limit to the length of commands that SQL*Plus can process in this version? My code runs fine on 10g and 11g databases/clients, whether it's on one line or two. Anyone know of a bug in 9i regarding this? And if there's a workaround (other than the obvious one of upgrading, of course!)

As mentioned, it's choking at the 258th character, not the 240th. Also, it's a SELECT statement, not a COPY command (one relates to the database, the other is an internal SQL*Plus feature). And someone else also kindly pointed out that SQL*Plus has a command limit length (i.e., database-related commands) of 2500 characters. So I don't think that's the issue.
I can be a bit more specific about this one now, though. I've a sequence of Solaris boxes running 9.2.0.1 up, and the results are that 9.2.0.1 is affected; and it's a problem for 9.2.0.7:
SQL> select * from v$version;
BANNER
Oracle9i Enterprise Edition Release 9.2.0.1.0 - 64bit Production
PL/SQL Release 9.2.0.1.0 - Production
CORE 9.2.0.1.0 Production
TNS for Solaris: Version 9.2.0.1.0 - Production
NLSRTL Version 9.2.0.1.0 - Production
SQL> select 'update myauditing_table set test_details='''||grantee||''' where myaudit_refno=''A.FE'';' from (select case when upper(value) in ('DBO','TRUE') then (select agg_concat(grantee) from dba_tab_privs where table_name='AUD$') else (select 'AUDIT_TRAIL is set to: '||value from dual) end as grantee from v$parameter where upper(name)='AUDIT_TRAIL' order by grantee);
SP2-0734: unknown command beginning "s set to: ..." - rest of line ignored.And...
SQL> select * from v$version;
BANNER
Oracle9i Enterprise Edition Release 9.2.0.7.0 - 64bit Production
PL/SQL Release 9.2.0.7.0 - Production
CORE 9.2.0.7.0 Production
TNS for Solaris: Version 9.2.0.7.0 - Production
NLSRTL Version 9.2.0.7.0 - Production
SQL> select 'update myauditing_table set test_details='''||grantee||''' where myaudit_refno=''A.FE'';' from (select case when upper(value) in ('DBO','TRUE') then (select agg_concat(grantee) from dba_tab_privs where table_name='AUD$') else (select 'AUDIT_TRAIL is set to: '||value from dual) end as grantee from v$parameter where upper(name)='AUDIT_TRAIL' order by grantee);
SP2-0734: unknown command beginning "s set to: ..." - rest of line ignored.But in 10.2.0.1, the problem disappears.
I am not sure, but it is perhaps related to metalink note 285913.1, since we did get some ORA-00600: internal error code, arguments: [qernsRowP], [1] errors in the alert log with a slightly different version of the above query. Changing the query obviously alters the test, though, so it may be completely unrelated.
Either way, I think it's definitely a version-specific SQL*Plus bug (though it would be nice if any other 9i owners out there could try it and report back!)
Anyway, the workaround is simply to submit the command on two separate lines, wherever possible. And of course... to upgrade.

SQL Plus: retrieve the commands

Similar Messages

Maybe you are looking for