I'm having this odd problem since yesterday. I've tried several options and I actually reinstalled ORACLE and the DB itself.
Here's the problem: I have this table that is somekind of zombie. Here are the symptoms:
SELECT TABLE_NAME FROM USER_TABLES WHERE TABLE_NAME='MYTABLE'
Returns a record, meaning that the table exists.
SELECT COLUMN_NAME FROM USER_TAB_COLUMNS WHERE TABLE_NAME = 'MYTABLE'
Returns all the columns of MYTABLE. So far so good, the table exists.
SELECT * FROM MYTABLE
Returns ORA-00942: table or view does not exist.
At this point I'm quite confused: the table seems to exist on the USERTABLES but I cannot SELECT over it?
CREATE TABLE MYTABLE (Foo NUMBER) TABLESPACE MYTABLESPACE
Returns:
ORA-00604: error occurred at recursive SQL level 1
ORA-00001: unique constraint (SYS.I_OBJ2) violated
I do not understand this error. But the best is yet to come.
SELECT * FROM MYTABLE
Surprisingly, the above query (an exact copy of the 3rd query) returns several records now!
Moreover, I noticed that the column Foo is not present: the table I now see is my initial table that had other columns.
DROP TABLE MYTABLE
I now try to drop the table and I get the following errors:
ORA-00604: error occurred at recursive SQL level 1
ORA-00942: table or view does not exist
ORA-06512: at line 19
SELECT * FROM MYTABLE
More confused than ever, I try the above query and, surprise surprise, the table no longer exists.
I don't undestand this: the table is on USERTABLES but I cannot SELECT over it, however, if I create a new table with the same name, I get an error but now I can SELECT over the previous version of that table with several records.
Any thoughts ? I really need your help :(
EDIT - I checked now: I'm unable to drop ANY table. This might just be a new symptom.
Solution
The problem was that MDSYS.SDO_GEOR_SYSDATA_TABLE table was missing and a drop event trigger was trying to access it, generating the error. The solution was restoring that table.
If have privileges, try this query:
SELECT *
FROM dba_objects
WHERE object_name = 'MYTABLE';
And see what objects exist with that name. It might point you in the right direction.
You didn't qualify the schema names when trying to select and drop. The CURRENT_SCHEMA of your session may be different form the log-on user. Check by trying
select SYS_CONTEXT('USERENV', 'CURRENT_SCHEMA') from dual;
Instead of describing what the output was, could you please copy/paste the complete output for us?
Lastly, can you exclude that someone messed up the dictionary? You know, SYSDBA can do anything....
Related
I've just gotten a new query error that I haven't changed anything to. Any advice on what to do? Thanks
SQL compilation error:
View definition for '**********' declared 115 column(s), but view query produces 117 column(s).
This is speculation, but it sounds like your view is using select x.*, where * means to get all the columns from some table.
Then, the underlying table changes . . . and voila, you might have a problem.
I've just gotten a new query error that I haven't changed anything to. Any advice on what to do?
If the query started to produce errors it means that the defintion of view is no longer "valid/up-to-date". Most likely the underlying table has been altered.
CREATE VIEW
View definitions are not dynamic. A view is not automatically updated if the underlying sources are modified such that they no longer match the view definition, particularly when columns are dropped. For example:
A view is created referencing a specific column in a source table and the column is subsequently dropped from the table.
A view is created using SELECT * from a table and any column is subsequently dropped from the table.
In either of these scenarios, querying the view returns a column mismatch error.
Steps to reproduce the scenario:
CREATE OR REPLACE TABLE t(col1 INT, col2 INT);
INSERT INTO t(col1, col2) VALUES (1,1);
CREATE OR REPLACE VIEW v_t AS SELECT * FROM t;
SELECT * FROM v_t;
--COL1 COL2
--1 1
So far so good. Now altering the underlying table by adding new column:
ALTER TABLE t ADD COLUMN col3 INT DEFAULT 3;
SELECT * FROM v_t;
SQL compilation error: View definition for 'V_T' declared 2 column(s), but view query produces 3 column(s).
Recreation of the view and keeping its definition on par with underlying tables should resolve it:
CREATE OR REPLACE VIEW v_t
COPY GRANTS
AS
SELECT * FROM t;
-- using * will work to refresh it but I would not recommend it
-- and explicitly describe columns instead
CREATE OR REPLACE VIEW v_t
COPY GRANTS -- to preserve already granted permissions
AS
SELECT col1, col2, col3 FROM t;
SELECT * FROM v_t;
-- COL1 COL2 COL3
-- 1 1 3
I found myself with a similar issue this morning. I had copied my query from a txt file I had and pasted it into a worksheet and tired to run it and got the same error. I had used the Table with the definition issue for a join and only for 1 column so I didn't see why it was giving me such an error.
All a look of check tables I commented on the Worksheet the error I was seeing. But I decided to run it again and it worked.
This tells me that Snowflake was using what it had cached but after editing the query it saw it as a new query and re-ran it instead of erroring out when what is in the cache doesn't match the definition.
I've been wrestling with setting up a trigger and keep getting the error:
SQL logic error near "DROP": syntax error
I have several tables main_table, other_one, other_two, etc.
main_table has several columns with the primary key column named filehash
The values in the primary key column of main_table are also the names of the other_* tables
So, if I delete a row in main_table with a primary key of other_one, I want the trigger to DROP the table other_one too
Here's the trigger statement that is producing the error
CREATE TRIGGER remove_other_one AFTER DELETE ON 'main_table'
WHEN (OLD.filehash == 'other_one')
BEGIN
DROP TABLE IF EXISTS 'other_one' ;
END remove_other_one;
EDIT: the 'fuller' error I get when I run the trigger statement in SQLite DB Browser is:
near "DROP": syntax error: CREATE TRIGGER remove_other_one AFTER DELETE ON 'main_table' WHEN (OLD.filehash == 'other_one') BEGIN DROP
Based on SQLite trigger doc I believe that it is not possible:
There is no option for DDL/dynamic SQL inside trigger.
I guess that you wanted to achieve something like PostgreSQL DBFiddle Demo 1 and Demo 2
You could handle your case in application code. Anyway table per date/customer/hash almost always indicates poor design and in long run will cause more problems.
I´m trying to execute stored procedure but I get an issue of an existing temporal table, but I just create one time and use into another part of code
SELECT ...
INTO #tmpUnidadesPresupuestadas
FROM proce.table1
--Insertar in table src..
INSERT INTO table (
....)
SELECT
....
FROM
#tmpUnidadesPresupuestadas
I get this message:
There is already an object named
'#tmpUnidadesPresupuestadas' in the database.
How can I solve it? Regards
A temp table lives for the entirety of the current session. If you run this statement more than once, then the table will already be there. Either detect that and truncate it, or before selecting into it drop it if it exists:
DROP TABLE IF EXISTS #tmpUnidadesPresupuestadas
If prior to SQL Server 2016, then you drop as such:
IF OBJECT_ID('tempdb.dbo.#tmpUnidadesPresupuestadas', 'U') IS NOT NULL
DROP TABLE #tmpUnidadesPresupuestadas;
Without seeing more of the code, it's not possible to know if the following situation is your problem, but it could be.
When you have mutually exclusive branches of code that both do a SELECT...INTO to the same temp table, a flaw causes this error. SELECT...INTO to a temp table creates the table with the structure of the query used to fill it. The parser assumes if that occurs twice, it is a mistake, since you can't recreate the structure of the table once it already has data.
if #Debug=1
select * into #MyTemp from MyTable;
else
select * into #MyTemp from MyTable;
While obviously not terribly meaningful, this alone will show the problem. The two paths are mutually exclusive, but the parser thinks they may both get executed, and issues the fatal error. You extend that, wrapping each branch in a BEGIN...END, and add the drop table (conditional or not) and the parser will still give the error.
To be fair, in fact both paths COULD be executed, if there were a loop or GOTO so that one time around #Debug = 1, and the other time it does not, so it may be asking too much of a parser. Unfortunately, I don't know of a workaround, and using INSERT INTO instead of SELECT INTO is the only way I know to avoid the problem, even though that can be terribly onerous to name all the columns in a particularly column-heavy query.
I am a bit unclear as to what you are attempting. I assume you don't want to drop the table at this point. I believe the syntax you may be looking for is
Insert Into
Insert into #tmpUnidadesPresupuestadas (Col1, col2, ... colN)
Select firstcol, secondcol... nthCol
From Data
If you do indeed wish to drop the table, the previous answers have that covered.
This might be useful for someone else, keep in mind that If more than one temporary table is created inside a single stored procedure or batch, they must have different names. If you use the same name you won't be able to ALTER the PROCEDURE.
https://learn.microsoft.com/en-us/previous-versions/sql/sql-server-2012/ms174979(v=sql.110)#temporary-tables
Make sure the stored procedure and the table doesn't have same name.
Add logic to delete if exists. Most likely you ran it previously. The table remains from the previous running of the stored procedure. If you log out and log in then run it, that would likely clear it. But the cleanest way is to check if it exists and delete it if it does. I assume this is MsSql.
At first you should check if temp table is already exist if yes then delete it then create a empty table then use insert statement. refer below example.
IF OBJECT_ID('tempdb..#TmpTBL') IS NOT NULL
DROP TABLE #TmpTBL;
SELECT TOP(0) Name , Address,PhoneNumber
INTO #TmpTBL
FROM EmpDetail
if #Condition=1
INSERT INTO #TmpTBL (Name , Address,PhoneNumber)
SELECT Name , Address,PhoneNumber FROM EmpDetail;
else
INSERT INTO #TmpTBL (Name , Address,PhoneNumber)
SELECT Name , Address,PhoneNumber FROM EmpDetail;
I want to write a procedure in SQL anywhere which can check if a local temporary table exists and if it does use it. I do not want to drop the table. I have already found a way to drop local temporary table which is:
DROP TABLE IF EXISTS t;
I have also tried following:
I created a local temporary table TEMP_TABLE. Then I tried to run this query:
select object_id('tempdb..TEMP_TABLE')
This just gives me NULL. But if I try
select * from TEMP_TABLE
it works perfectly fine.
So can anyone please help me find a way to check if the local temporary table exists in sql anywhere.
I'm not sure what version of Sybase you have but this works in Sybase 11 so I can imagine it will work in any version up too:
Begin
Create local Temporary table TEMP_TABLE (column1 int); //Create temp table
// any other code needed to be executed if table did not exist
Exception when others then
// Code to be executed when table does exist
end;
This is basically a try..catch for sybase. If the Temp Table exists it will throw an exception, in the exception you can run the code you want to knowing that the table already exists.
In one query you are referring database and in other you are not, try below two queries.
select object_id('tempdb..TEMP_TABLE')
select * from tempdb..TEMP_TABLE
select object_id('TEMP_TABLE')
select * from TEMP_TABLE
I know the statement:
create table xyz_new as select * from xyz;
Which copies the structure and the data, but what if I just want the structure?
Just use a where clause that won't select any rows:
create table xyz_new as select * from xyz where 1=0;
Limitations
The following things will not be copied to the new table:
sequences
triggers
indexes
some constraints may not be copied
materialized view logs
This also does not handle partitions
I used the method that you accepted a lot, but as someone pointed out it doesn't duplicate constraints (except for NOT NULL, I think).
A more advanced method if you want to duplicate the full structure is:
SET LONG 5000
SELECT dbms_metadata.get_ddl( 'TABLE', 'MY_TABLE_NAME' ) FROM DUAL;
This will give you the full create statement text which you can modify as you wish for creating the new table. You would have to change the names of the table and all constraints of course.
(You could also do this in older versions using EXP/IMP, but it's much easier now.)
Edited to add
If the table you are after is in a different schema:
SELECT dbms_metadata.get_ddl( 'TABLE', 'MY_TABLE_NAME', 'OTHER_SCHEMA_NAME' ) FROM DUAL;
create table xyz_new as select * from xyz where rownum = -1;
To avoid iterate again and again and insert nothing based on the condition where 1=2
Using sql developer select the table and click on the DDL tab
You can use that code to create a new table with no data when you run it in a sql worksheet
sqldeveloper is a free to use app from oracle.
If the table has sequences or triggers the ddl will sometimes generate those for you too. You just have to be careful what order you make them in and know when to turn the triggers on or off.
You can do this
Create table New_table as select * from Old_table where 1=2 ;
but be careful
The table you create does not have any Index, PK and so on like the old_table.
DECLARE
l_ddl VARCHAR2 (32767);
BEGIN
l_ddl := REPLACE (
REPLACE (
DBMS_LOB.SUBSTR (DBMS_METADATA.get_ddl ('TABLE', 'ACTIVITY_LOG', 'OLDSCHEMA'))
, q'["OLDSCHEMA"]'
, q'["NEWSCHEMA"]'
)
, q'["OLDTABLSPACE"]'
, q'["NEWTABLESPACE"]'
);
EXECUTE IMMEDIATE l_ddl;
END;
Simply write a query like:
create table new_table as select * from old_table where 1=2;
where new_table is the name of the new table that you want to create and old_table is the name of the existing table whose structure you want to copy, this will copy only structure.
SELECT * INTO newtable
FROM oldtable
WHERE 1 = 0;
Create a new, empty table using the schema of another. Just add a WHERE clause that causes the query to return no data:
WHERE 1 = 0 or similar false conditions work, but I dislike how they look. Marginally cleaner code for Oracle 12c+ IMHO is
CREATE TABLE bar AS
SELECT *
FROM foo
FETCH FIRST 0 ROWS ONLY;
Same limitations apply: only column definitions and their nullability are copied into a new table.
If one needs to create a table (with an empty structure) just to EXCHANGE PARTITION, it is best to use the "..FOR EXCHANGE.." clause. It's available only from Oracle version 12.2 onwards though.
CREATE TABLE t1_temp FOR EXCHANGE WITH TABLE t1;
This addresses 'ORA-14097' during the 'exchange partition' seamlessly if table structures are not exactly copied by normal CTAS operation. I have seen Oracle missing some of the "DEFAULT" column and "HIDDEN" columns definitions from the original table.
ORA-14097: column type or size mismatch in ALTER TABLE EXCHANGE
PARTITION
See this for further read...
you can also do a
create table abc_new as select * from abc;
then truncate the table abc_new. Hope this will suffice your requirement.
Using pl/sql developer you can right click on the table_name either in the sql workspace or in the object explorer, than click on "view" and than click "view sql" which generates the sql script to create the table along with all the constraints, indexes, partitions etc..
Next you run the script using the new_table_name
copy without table data
create table <target_table> as select * from <source_table> where 1=2;
copy with table data
create table <target_table> as select * from <source_table>;
In other way you can get ddl of table creation from command listed below, and execute the creation.
SELECT DBMS_METADATA.GET_DDL('TYPE','OBJECT_NAME','DATA_BASE_USER') TEXT FROM DUAL
TYPE is TABLE,PROCEDURE etc.
With this command you can get majority of ddl from database objects.
Create table target_table
As
Select *
from source_table
where 1=2;
Source_table is the table u wanna copy the structure of.
create table xyz_new as select * from xyz;
-- This will create table and copy all data.
delete from xyz_new;
-- This will have same table structure but all data copied will be deleted.
If you want to overcome the limitations specified by answer:
How can I create a copy of an Oracle table without copying the data?
The task above can be completed in two simple steps.
STEP 1:
CREATE table new_table_name AS(Select * from old_table_name);
The query above creates a duplicate of a table (with contents as well).
To get the structure, delete the contents of the table using.
STEP 2:
DELETE * FROM new_table_name.
Hope this solves your problem. And thanks to the earlier posts. Gave me a lot of insight.