I have two schemas: A and B (Oracle 9). At the A there is a dblink to B. At the B there is a package, that i calls from A. Procedures in B package can returns varying count results and i think that returning a collection is a better way for this reason.
create type B.tr_rad as object (
name varchar2(64)
,code number
,vendor number
,val varchar2(255)
,num number
);
create type B.tt_rad as varray(256) of B.tr_rad;
But from A scheme I cannot use tt_rad type because using SQL-types by dblink is not supported. DBMS_SQL is not supported cursors. Create types with same OID is impossible.
I think to use temporary tables. But firstly it is not that good (after the remote function returns the value, calling side must select collection from remote table). And there are fears of a slowdown of work with temporary tables.
Maybe who knows the alternative interaction?
I've had similar problems in the past. Then I came to the conclusion that fundamentally Oracle's db links are "broken" for anything but simple SQL types (especially UDT's, CLOBS may have problems, XMLType may as well). If you can get the OID solution working then good luck to you.
The solution I resorted to was to use a Java Stored procedure, instead of the DB Link.
Characteristics of the Java Stored Procedure:
Can return a "rich set of types", just about all of the complex types (UDT's, tables/arrays/varrays) see Oracle online documentation for details. Oracle does a much better job of marshalling complex (or rich) types from java, than from a DBLink.
Stored Java can acquire the "default connection" (runs in the same session as the SQL connection to the db - no authentication issues).
Stored Java calls the PL/SQL proc on the remote DB, and the java JDBC layer does the marshaling from the remote DB.
Stored Java packages up the result and returns the results to the SQL or PL/SQL layer.
It's a bit of work, but if you have a bit of java, you should be able to "cut and paste" a solution together from the Oracle documentation and sample.
I hope this helps.
See this existing discussion
referencing oracle user defined types over dblink
An alternative interaction is to have one database with schemas A and B instead of two databases with a database link.
My solution.
On the side B i create temporary table like the collection record. At the A side i have a DBMS_SQL wrapper that calls procedure over dblink. This procedure writes result collection in the temporary table. After successful completion remote procedure i select results from remote temporary table and transform it to local collection type.
Limitations
1. the need for permanent object synchronization.
2. impossibility use A-side procedure (that call remote procedure) in SQL query.
3. the complexity of using.
Related
I am not sure in case of Stored Procedures, if Postgresql treats static sql any differently from a query submitted as a quoted string.
When I create a stored procedure in PostgreSQL using static sql, there seems to be no validation of the table names and table columns or column types but when I run the procedure I get the listing of the problems if any.
open ref_cursor_variable for
select usr_name from usres_master;
-- This is a typing mistake. The table name should be users_master. But the stored procedure is created and the error is thrown only when I run the procedure.
When I run the procedure I (naturally) get some error like :
table usres_master - invalid table name
The above is a trivial version. The real procedures we use at work combine several tables and run to at least a few hundred lines. In PostgresQL stored procedure, is there no advantage to using static sql over dynamic sql i.e. something like open ref_cursor_variable for EXECUTE select_query_string_variable.
The static SQL should be preferred almost time - dynamic SQL should be used only when it is necessary
from performance reasons (dynamic SQL doesn't reuse execution plans). One shot plan can be better some times (and necessary).
can reduce lot of code
In other cases uses static SQL every time. Benefits:
readability
reuse of execution plans
it is safe against SQL injection by default
static check is available
The source of a function is just a string to Postgres. The main reason for this is the fact that Postgres (unlike other DBMS) supports many, even installable languages for functions and procedures. As the Postgres core can't possibly know the syntax of all languages, it can not validate the "inner" part of a function. To my knowledge the "language API" does not contain any "validate" method (in theory this would probably be possible though).
If you want to statically validate your PL/pgSQL functions (and procedures since Postgres 11) you could use e.g. https://github.com/okbob/plpgsql_check/
First of all usually I am working with MSSQL. But I have a stored procedure in MSSQL, which I need to use in Oracle now and since I am absolutely new to Oracle I have no idea at all how to do it correct.
I needed to use user defined table types in my MS SQL stored procedure because I am using "logical" tables in my stored procedure, which I also need to pass them to a dynamic sql statement within this procedure (using column names of "physical" tables as variables/parameters).
I've started to add the oracle function in a package I made before for another function. It looks like
TYPE resultRec IS RECORD
(
[result columns]
);
TYPE resultTable IS TABLE OF resultRec;
Function MyFunctionName([A LOT PARAMETERS]) RETURN resultTable PIPELINED;
I also described the layout of the tables (the user defined table types in MSSQL), which I want to use within this function in this package header.
So far so good, but now I don't really know where I have to declare my table variables or user defined table types. I also tried to put them in the package header, but if I am trying to use these tables in the package body, where I am describing my function, Oracle tells met, that the table or view does not exist.
I also tried it to describe the tables within the package body or in the block of my function, which looks like that:
FUNCTION MyFunctionName
(
[MyParameters]
)
RETURN resultTable PIPELINED is rec resultrec;
TYPE tableVariableA IS TABLE OF tableRecA;
TYPE tableVariableB IS TABLE OF tableRecB;
BEGIN
INSERT INTO tableVariableA
SELECT ColumnA, ColumnB FROM physicalTable WHERE[...];
[A LOT MORE TO DO...]
END;
But in this case Oracle also tells me, that it doesn't know the table or view.
I also tried a few more things, but at the end I wasn't able to tell Oracle what table it should use...
I would appreciate every hint, which helps me to understand how oracle works in this case. Thanks a lot!
You can't insert into a collection (e.g. PL/SQL table). You can use the bulk collect syntax to populate the collection:
SELECT ColumnA, ColumnB
BULK COLLECT INTO tableVariableA
FROM physicalTable
WHERE [...];
However, you might want to check this is an appropriate approach, since SQL Server and Oracle differ quite a bit. You can't use PL/SQL tables in plain SQL (at least prior to 12c), even inside your procedure, so you might need a schema-level type rather than a PL/SQL type, but it depends what you will do next. You might not really want a collection at all. Trying to convert T-SQL straight to PL/SQL without understanding the differences could lead you down a wrong path - make sure you understand the actual requirement and then find the best Oracle mechanism for that.
I'm used to Microsoft SQL Server, where the last SELECT query of a stored procedure determines what is returned.
In Oracle, I'm always using a SYS_REFCURSOR OUT parameter to return data from queries.
Are there other options of returning data from SELECT queries in Oracle stored procedures?
How about SELECT queries that select only one row? Is a SYS_REFCURSOR still necessary?
EDIT: I need to know the answer for Oracle 11g R2 (should have mentioned that explicitly instead of just in the tags).
Until now Oracle has not supported the SQL Server style of procedures implicitly returning a result set, you have had to explicitly return something - which could be a SYS_REFCURSOR or a collection or whatever.
In Oracle 12C a new feature has been added called Implicit Statement Results, which is designed to emulate the SQL Server way of working. However, this is really intended to support migration of existing code from SQL Server; for fresh Oracle development you would be best advised to learn the way Oracle normally does things.
Oracle PL/SQL Procedures can return all the supported basic datattype( date,varchar2,number) plus complex ( records, tables, varray)
Another option is the pipelined function, in which you call the function as:
select ...
from table(my_function(param1 => 1, ...))
See docs for details.
Just for clarifying, ORACLE procedures cannot 'RETURN" per se, a SYS_REFCURSOR OUT parameter is more like changing the value of a variable reference inside the procedure.
Apart from SYS_REFCURSOR, if you are returning only one row of a table, say EMPLOYEE, you can also define a record as EMPLOYEE%ROWTYPE and use it as a OUT type.
Or like:
PROCEDURE pr_proc (v_input in number
v_emp_row out EMPLOYEE%ROWTYPE )
IS
I just created a stored procedure in MS SQL DB using TOAD.
what it does is that it accepts an ID wherein some records are associated with, then it inserts those records to a table.
next part of the stored procedure is to use the ID input to search on the table where the items got inserted and then return it as the result set to the user just to confirm that the information got inserted.
IN TOAD, it does what is expected. It inserts date and returns information using just the stored procedure.
IN Oracle SQL developer however, it does the insert and it ends at that. It seems to not execute the 2nd part of the stored procedure which is a select stmt.
I just have a feeling that this is because of the jdbc adapter. Also why I'm asking is because I'm using a reporting tool Pentaho Report Designer and it would really make it easier if I can do 2 things at the same time. Pentaho Report Designer is also using jdbc adapters, not a coincidence maybe?
But if there are other things that I can tweak I'd really appreciate it.
This is a guess, but worth considering...
There are things called "Batches", where are sets of SQL Statements that are all sent to the server at once, and executed by the server as one set of statements, within a single server-side session. Sending a set of sql statements to the server as a batch will often result in different results than if you sent them one at a time, where each statement is executed in its own session.
I haven't used Toad (or Oracle) in a while, but as I recall, it dealt with batches differently than the other ide I used. If the second statement in your set is relying on being in the same session as the first, and in one ide it is in a separate session, then this might explain what is happening.
We have an application written in Delphi 2010 which connects to SQL Server Database. Now we're in the process of migrating to Oracle. With SQL Server it was very easy to perform insert, update, delete right from a dbgrid connected to a Stored Procedure.
It's because stored procedures in SQL Server can easily act as a table so that you can do any operation on it, providing it returns the necessary columns within the resultset. Now with Oracle I don't know how do do it. I connect a DBGrid to a DataSource, dataset of which is a Stored Procedure object,but I can't edit the grid. Just Select is possible.
What do I have to do to to achieve this?I use UniDac component suite to connect to Oracle database.
Oracle does not support such functionality. IOW, in Oracle you cannot edit result set provided by a stored procedure or include stored procedure into INSERT INTO <name>, UPDATE <name> or DELETE FROM <name>.
While it is traditional for SQL Server developers to "always" use stored procedures (due to many reasons), it is not traditional for Oracle developers. But it is possible with Oracle too. Search for "REF CURSOR" to see how to fetch data using SP. And use normal or packaged (preferred) SP to post updates to a DB. These procedure will receive old / new field values through arguments.
I cannot say precisely about UniDAC, I can say about AnyDAC. But I will expect UniDAC has similar functionality. To use SP for posting updates you will need to use TXxxUpdateSQL component.
OK,here I'm answering the question though I can see very few are dealing with Delphi recently. Let's say we have a stored proc in Oracle database:
CREATE OR REPLACE PROCEDURE GET_EMPLOYEES
(V_CUR IN OUT SYS_REFCURSOR)
AS
BEGIN
OPEN V_CUR FOR SELECT * FROM EMPLOYEES;
END GET_EMPLOYEES;
Now, in Delphi you pick a stored procedure component (probably from ODAC or UniDac component suite).Set its StoredProcName GET_EMPLOYEES. Then you can add all the fields that the procedure returns in a cursor.If you run the application and activate the stored procedure you'll be able to see all the records. But if you try to insert, modify or delete anything you'll fail to do so. Now, there's a very tricky thing. If you check, you'll see that ReadOnly property of all fields are set to True. Even after you set them to False nothing will change in the real database, although you can edit the DBGrid.
So, we've come to the main part. How did the old Delphi-SQL Server partnership work so that you could do any operation right from a DBGrid? Well, we must understand that there's no magic. If it's SQL, then SQL has only one way of INSERTING,UPDATING and DELETING records-it's with the appropriate SQL statements.With Delphi-SQL Server there seems to be an implicit SQL statement that we never paid attention. But with Oracle, we have to provide our own statements for each operation.
If you use UniDac or ODAC then there's SQLInsert,SQLUpdate,SQLDelete properties in a StoredProc object.If you want to insert a record through DBGrid, then you should edit its SQLInsert property to
INSERT INTO EMPLOYEES VALUES(:EMPLOYEEID,:EMPLOYEENAME)
where variables following : are corresponding to te fields of the stored procedure.They're simply bind variales.When updating and deleting though you'll need some unique value to represent a specific record. Primary key is one option(maybe the only option as I haven't been able to figure out how to use ROWID for the same purpose).So the sql statements for UPDATE and DELETE would be
DELETE FROM EMPLOYEES WHERE EMPLOYEEID=:EMPLOYEEID
and
UPDATE EMPLOYEES SET EMPLOYEENAME=:EMPLOYEENAME WHERE EMPLOYEEID=:EMPLOYEEID
P.S. I just found a way to use ROWID for update and delete statements. In your stored procedure if you choose ROWID too and give it an alias then you can construct your UPDATE and DELETE Statements like such:
UPDATE EMPLOYEES SET EMPLOYEENAME=:EMPLOYEENAME,..... WHERE ROWID=:RECORD_ROWID
DELETE FROM EMPLOYEES WHERE ROWID=:RECORD_ROWID
In the preceding statements RECORD_ROWID is the fieldname returned from stored procedure as a result of aliasing ROWID. If you use :ROWID instead you'll get "ORA-01745: invalid host/bind variable name" error. This is because in a binding variable a colon cannot be followed by a reserved word. And ROWID is a reserved word.