Why the pipelined function working like normal table function? - sql

I have created one pipelined function and another non-pipelined function.
But both functions when called from select statement, all results are shown only after complete execution of loop.
Why the pipelined function is not returning value as soon as the data for the each row is ready?
SELECT * FROM TABLE(GET_TAB(10,1));--Normal function call
SELECT * FROM TABLE(GET_TAB_P(10,1));--Pipelined function call
create or replace TYPE T_TF_ROW AS OBJECT(CNT NUMBER, DESCRIPTION VARCHAR2(50));
create or replace TYPE T_TF_TAB AS TABLE OF T_TF_ROW;
create or replace FUNCTION GET_TAB(P_ROWS IN NUMBER, P_SLEEP IN NUMBER)
RETURN T_TF_TAB
AS
V_T_TF_TAB T_TF_TAB:=T_TF_TAB();
BEGIN
FOR I IN 1..P_ROWS LOOP
DBMS_LOCK.SLEEP(P_SLEEP);
V_T_TF_TAB.EXTEND;
V_T_TF_TAB(V_T_TF_TAB.LAST):=T_TF_ROW(I,'DESCRIPTION OF : '||I);
END LOOP;
RETURN V_T_TF_TAB;
END;
create or replace FUNCTION GET_TAB_P(P_ROWS IN NUMBER, P_SLEEP IN NUMBER)
RETURN T_TF_TAB PIPELINED
AS
BEGIN
FOR I IN 1..P_ROWS LOOP
DBMS_LOCK.SLEEP(P_SLEEP);
PIPE ROW(T_TF_ROW(I,'DESCRIPTION FOR ' || I));
END LOOP;
RETURN;
END;
Normal function result
Pipelined function result

Your pipelined function is working correctly, but your client's fetch size optimization is retrieving N rows at-a-time before it returns anything. Disable or lower the fetch size and you should start seeing individual rows returned. However, in production you do not want to disable that optimization.
In SQL*Plus, you would run this command to see one-row-at-a-time:
set arraysize 1;
(That actually returns two rows at a time for a reason I can't remember. There is a workaround, but unless you are writing some sort of progress-bar program you probably don't care about getting exactly one row-at-a-time.)

Related

How to find what data caused Oracle function failure?

sorry if this not the right place.
I am doing a SQL SELECT statement, invoking a function. It's a large data dump - about 10,000 records.
I am calling a function to preform some calculations, but its failing.
One ore more of those records has bad data that is causing the function to crash.
Is there any way to see exactly what data caused the crash readily Or should I create some code to run the function by hand for each of 10,000 records? I could create code that generates the input data fairly straightforwardly, then run the function like this SELECT MY_FUNCT(1,1,1) FROM DUAL; but I am wondering if there is a better way.
For reference I am running the SQL query like this.
SELECT
MY_FUNCT(A.FOO, A.BAR)
FROM TABLE A
WHERE ....;
As others have said, you just need to handle the error and not raise it all the way. A neat way of doing this would be to create a wrapper function for your function that sometimes fails, you can declare this function within your select query using a with pl/sql clause:
Let's say this is your function that sometimes fails
create or replace function my_funct (inputnumber number)
return varchar2
is
sString varchar2(200);
begin
if inputnumber = 42 then
raise_application_error(-20001,'UH OH!');
end if;
sString := 'Input: '||inputnumber;
return sString;
end my_funct;
/
We can define a function that takes the same inputs, and just calls this function, then we just need to add some error handling (obviously never just rely on dbms_output to capture errors, this is just to make it obvious):
function my_funct_handle (inputnumber number)
return varchar2
is
begin
return my_funct (inputnumber => inputnumber);
exception when others then
dbms_output.put_line(sqlerrm||' at '||inputnumber);
return 'ERROR';
end;
And we can then just stick that in our query using with function
with
function my_funct_handler (inputnumber number)
return varchar2
is
begin
return my_funct (inputnumber => inputnumber);
exception when others then
dbms_output.put_line(sqlerrm||' at '||inputnumber);
return 'ERROR';
end;
select my_funct_handler (id), string_col
from as_table;
/
I get both the dbms_output text to describe the error and the ID but I could also filter on the results of that function to only show me the erroring rows:
with
function my_funct_handle (inputnumber number)
return varchar2
is
begin
return my_funct (inputnumber => inputnumber);
exception when others then
dbms_output.put_line(sqlerrm||' at '||inputnumber);
return 'ERROR';
end;
select my_funct_handle (id), string_col
from as_table
where my_funct_handle (id) = 'ERROR';
/
MY_FUNCT_HANDLE(ID)
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
STRI
----
ERROR
blah
ORA-20001: UH OH! at 42
ORA-20001: UH OH! at 42
(I get two errors shown in the dbms_output output as the function can be called multiple times - once as part of the select list and once as part of the where clause.)
One option is to handle the exception properly, e.g.
create or replace function my_funct(par_foo in number, par_bar in number)
return number
is
retval number;
begin
select sal
into retval
from emp
where ename = par_foo
and deptno = par_bar;
return par_foo/par_bar;
exception --> this
when no_data_found then
return null;
end;
If you want, you can even log those errors. How? Make the function autonomous transaction (so that it could write into the log table; you'll have to commit that insert). Store all relevant information (including SQLERRM). Once your code finishes, check what's written in the log file and then decide what to do.
Or, you could even continue current task by enclosing that select into its own begin-exception-end block within a loop, e.g.
begin
for cur_r in (select ... from ...) loop
begin
-- your current SELECT
SELECT
MY_FUNCT(A.FOO, A.BAR)
FROM TABLE A
WHERE ....;
exception
when others then
dbms_output.put_line(cur_r.some_value ||': '|| sqlerrm);
end;
end loop;
end;
one better approach is to create a error handler Package/Procedure which will write it into a table and call it from the function, this way all the errors will be captured in a Oracle table.
--- untested -- You can create other columns to capture the function name, date, and other details in the error table.
PROCEDURE SP_ERROR_INS_ERR_COMMON (n_ERR_CODE NUMBER, c_ERR_SOURCE VARCHAR2, n_ERR_LINE NUMBER, c_ERR_DESC VARCHAR2, C_USER_COMMENT VARCHAR2 ) IS
n_Log_Id NUMBER;
PRAGMA AUTONOMOUS_TRANSACTION;
BEGIN
n_Log_Id := 0;
INSERT INTO ERROR_LOG_COMMON
(ERROR_CODE, ERROR_SOURCE, ERROR_LINE, ERROR_DESCRIPTION, USER_COMMENT)
VALUES
(n_ERR_CODE, c_ERR_SOURCE, n_ERR_LINE, c_ERR_DESC, C_USER_COMMENT
);
COMMIT;
raise_application_error( -20001,SUBSTR(SQLERRM, 1, 200));
END SP_ERROR_INS_ERR_COMMON;
In your function you can call the error
EXCEPTION
WHEN OTHERS
THEN
vn_errcode := SQLCODE;
vc_errmsg := SUBSTR (SQLERRM, 1, 4000);
sp_error_ins_err_common(vn_ErrCode,
'SP_RP_DIM_COMPARE', NULL, vc_ErrMsg, substr('batch_id ' || g_BATCH_ID ||
l_str_err_msg,1,4000) );
RAISE;

How to read multiple refcursor return by other procedure to another procedure

i'm having one procedure which returns setof cursors
Now i have to call that procedure to another procedure and access the data
that return by that procedure
is their any way to do this in postgres.
This is code for 1st procedure,
CREATE OR REPLACE FUNCTION public.returns_multiple_cursor( )
RETURNS SETOF refcursor
LANGUAGE 'plpgsql'
COST 100.0
AS $function$
DECLARE
_daily refcursor := 'first_cur';
_fac_hourly refcursor := 'second_cur';
BEGIN
open first_cur for
select * from x;
return next first_cur;
open second_cur for
select * from y;
return second_cur;
END
$function$;
ALTER FUNCTION public.returns_multiple_cursor();
Here code for other second procedure
CREATE OR REPLACE FUNCTION public.access_cursor( )
RETURNS SETOF refcursor
LANGUAGE 'plpgsql'
COST 100.0
AS $function$
DECLARE
BEGIN
-- what code will be here to access the cursor data in this procedure
select public.returns_multiple_cursor();
END;
ALTER FUNCTION public.access_cursor();
Unfortunately, you cannot use the FOR <recordvar> IN <cursor> loop, because it only works for bound cursors (which refcursors are not).
But you can still loop through them, with the old-fashioned FETCH:
declare
rec record;
cur refcursor;
begin
for cur in select returns_multiple_cursor() loop
loop
fetch next from cur into rec;
exit when not found;
-- do whatever you want with the single rows here
end loop;
close cur;
end loop;
end
Unfortunately, there is still another limitation: PostgreSQL caches the first cursor's plan (at least, it seems it does something like that), so you must use cursors, which uses the same column types (you'll have to use the same column names anyway, to be able to refer them in the inner loop, like rec.col1).
Complete, working example: http://rextester.com/FNWG91106 (see f.ex. what happens, when you remove casting from the cursors' queries).
If you have fix number of cursors (like in your example), but differently structured underlying queries, it might be easier to declare your returns_multiple_cursor as:
create or replace function returns_multiple_cursor(out first_cur refcursor,
out second_cur refcursor)
-- returns record (optional)
language plpgsql
-- ...
This way, you could access your cursors more directly in the calling context.
Update: it seems that when you don't use explicit column names, just generic record processing (via f.ex. JSON or hstore), plan caching does not cause any trouble:
http://rextester.com/QHR6096

How to output multiple rows from an existing table in a stored procedure using oracle sql developer?

I have an existing Customers table and want to output each row from this table using a stored procedure. There is no input criteria, just need to output all of the records.
The stored procedure should basically be equivalent to:
"SELECT C_ID, LAST, FIRST, DOB, DPHONE, EMAIL FROM customers;"
This seems simple but I can't figure it out. All my searches haven't worked out for this case.
How would one accomplish this?
EDIT: Answered my question below. Very simple.
In Oracle your options are:
1. Use a function and return a REF CURSOR
2. Use a procedure and use a REF CURSOR as an OUT parameter
3. Use a PIPELINED function
4. Use a function and return a collection.
read this documentation
Overview of Table Functions
see similar question here Return collection from packaged function for use in select
simple sample of such function
CREATE FUNCTION StockPivot(p refcur_pkg.refcur_t) RETURN TickerTypeSet
PIPELINED IS
out_rec TickerType := TickerType(NULL,NULL,NULL);
in_rec p%ROWTYPE;
BEGIN
LOOP
FETCH p INTO in_rec;
EXIT WHEN p%NOTFOUND;
-- first row
out_rec.ticker := in_rec.Ticker;
out_rec.PriceType := 'O';
out_rec.price := in_rec.OpenPrice;
PIPE ROW(out_rec);
-- second row
out_rec.PriceType := 'C';
out_rec.Price := in_rec.ClosePrice;
PIPE ROW(out_rec);
END LOOP;
CLOSE p;
RETURN;
END;
/
Nevermind! Just got it. It is pretty simple. I just did this:
CREATE OR REPLACE PROCEDURE CustomerReport AS
BEGIN
FOR cus IN (SELECT C_ID, LAST, FIRST, DOB, DPHONE, EMAIL FROM customers)
LOOP
dbms_output.put_line(cus.C_ID||' '||cus.FIRST||' '||cus.LAST||' '||cus.DOB||' '||cus.DPHONE||' '||cus.EMAIL);
END LOOP;
END;

Procedures or Functions (Pl/sql)

I am currently learning pl/sql using oracle 10g
I have a certain confusion
When should I use stored procedures and when should i go for functions?
Please help me out with some real world example.
Thank you.
A function returns a value, although that could actually be an object like a cursor.
Also only a function can be defined with the following (as of 11.1):
DETERMINISTIC option, which helps the optimizer avoid redundant function calls.
PARALLEL_ENABLED option, which allows the function to be used safely in slave sessions of parallel DML evaluations.
PIPELINED option, which returns the results of a table function iteratively.
RESULT_CACHE option, which stores function results in the PL/SQL function result cache.
RESULT_CACHE clause, which specifies the data sources on which the results of a function.
So if you need to return multiple values, use a procedure. However be aware that the above five features are then not available.
If you want to include a PL/SQL subprogram in a SQL statement then you probably want a function.
http://docs.oracle.com/cd/B28359_01/appdev.111/b28370/subprograms.htm#CHDBEJGF
DECLARE
l_user_id VARCHAR2(1);
l_received_user VARCHAR2(30);
PROCEDURE print_user_name(user_name_in IN VARCHAR2)
AS
BEGIN
DBMS_OUTPUT.PUT_LINE('The user''s name is: ' || INITCAP(user_name_in));
END print_user_name;
FUNCTION get_user_name(user_id_in IN VARCHAR2) RETURN VARCHAR2
AS
l_user_name VARCHAR2(30);
BEGIN
SELECT 'Amanda'
INTO l_user_name
FROM dual
WHERE dummy = user_id_in;
RETURN l_user_name;
END get_user_name;
BEGIN
-- excute an action --
print_user_name('John');
l_user_id := 'X';
-- hold action's result in a variable --
l_received_user := get_user_name(l_user_id);
-- work with the received result/variable --
DBMS_OUTPUT.PUT_LINE('The received user''s name is: ' || INITCAP(l_received_user));
IF l_received_user = 'John' THEN
DBMS_OUTPUT.PUT_LINE('The received user''s name is John');
ELSE
DBMS_OUTPUT.PUT_LINE('The received user''s name is not John');
END IF;
END;
/*
The user's name is: John
The received user's name is: Amanda
The received user's name is not John
*/
Difference is that sored procedure do something, while functions do something and return result (variable or table).

PL SQL How to put many wrappers in one package

reference: How to make a wrapper to return something other than ref cursor
I've got many wrapped functions that are similar to the one below, only they return different columns of course. How can i put them all in one package. because what the examples below does is replace my wrapper package everytime.
I want to know how to remove the global declarations on top so that i can add many in the same package each with different return columns
create or replace package WrapperSample is
type TResultRow is record(
if_type codes.cd%type
,number_infected Integer);
type TResultRowList is table of TResultRow;
function GetADedIcWarningsProv
(
p_hos_id in work_entity_data.hos_id%type
,p_date in date
) return TResultRowList
pipelined;
end WrapperSample;
/
create or replace package body WrapperSample is
function GetADedIcWarningsProv
(
p_hos_id in work_entity_data.hos_id%type
,p_date in date
) return TResultRowList
pipelined is
v_refcur eOdatatypes_package.eOrefcur;
currentRow TResultRow;
begin
v_refcur := YourSchema.getADedIcWarningsProv(p_hos_id, p_date);
loop
fetch v_refcur
INTO currentRow;
exit when v_refcur%NotFound;
pipe row(currentRow);
end loop;
close v_refcur;
return;
end;
end WrapperSample;
/
you would use one record definition and one table definition per different column set.
create or replace package WrapperSample is
type R_WarningsProv is record(/*...*/);
type T_WarningsProv is table of R_WarningsProv ;
function GetADedIcWarningsProv(/*...*/) return T_WarningsProv pipelined;
type R_OtherFunction is record(/*...*/);
type T_OtherFunction is table of R_OtherFunction ;
function OtherFunction(/*...*/) return T_OtherFunction pipelined;
/* Multiple functions can use the same types as long as
they share the same column definition */
function SomeOtherFunction(/*...*/) return T_OtherFunction pipelined;
end WrapperSample;