I have a procedure like below:
SET serveroutput ON
DECLARE
TYPE list_of_names_t
IS TABLE OF emp.emp_index%TYPE;
ignoreIndexes LIST_OF_NAMES_T := List_of_names_t();
BEGIN
-- Logic here which fills the values in the collection ignoreIndexes and considerIndexes
-- Line XX
END;
At Line XX, I want to add the below line
SELECT * FROM emp WHERE emp_index NOT IN ignoreIndexes But not able to get the correct syntax, How to make that possible?
If the collection type is declared at the schema level, it can be used in SQL statements or dynamically .So do the following steps to make this work .
CREATE OR REPLACE TYPE list_of_names AS OBJECT(emp_index NUMBER);
CREATE OR REPLACE TYPE list_of_names_t IS TABLE OF list_of_names;
Create your anonymous block as shown below:
SET serveroutput ON
DECLARE
ignoreIndexes LIST_OF_NAMES_T := List_of_names_t();
emp_cur SYS_REFCURSOR;
BEGIN
-- Populate the collection with some data
SELECT list_of_names(LEVEL)
BULK COLLECT INTO ignoreIndexes
FROM dual CONNECT BY LEVEL<21;
OPEN emp_cur FOR
SELECT *
FROM EMP e
WHERE NOT EXISTS(SELECT 1
FROM table(ignoreIndexes) tt
WHERE tt.emp_index = e.emp_index
);
END;
Related
I am new to oracle.I already have a table tempash.
So,I created a procedure to see the data of this table.
So,I created procedure as:
create or replace procedure offc.temp_sel(data1 varchar2) is
var1 varchar2(4000);
BEGIN
var1:='select * from offc.temp'||data1;
EXECUTE IMMEDIATE var1;
end;
So,I executed the statement but,it is not showing me anything.
exec offc.temp_sel('ash');
There is no any compilation error in my procedure.But why the select statement is not showing me data of that procedure?
Try adding out parameter:
create or replace procedure offc.temp_sel(data1 varchar2,result out sys_refcursor)
is
BEGIN
open result for 'select * from offc.temp'||data1;
end;
SQL> var rc refcursor
SQL> execute offc.temp_sel('ash',:rc)
PL/SQL procedure successfully completed.
SQL> print rc
You need to keep the result of the SELECT statement into a variable. As you perform SELECT * ..., you should put the result into a RECORD type but, as the result set contains more than 1 row, your variable needs to be a table of records.
In order to not be prone to error, the tables of records needs to be exactly like your table source structure.
CREATE OR REPLACE PROCEDURE OFFC.TEMP_SEL(DATA1 VARCHAR2) IS
VAR1 VARCHAR2(4000);
TYPE T_RESULT IS TABLE OF offc.temp%ROWTYPE;
-- defined the new type based on the structure of table TEMP from schema OFFC
v_result t_result;
-- define a variable of that type.
BEGIN
var1:='select * from offc.temp'||data1;
EXECUTE IMMEDIATE VAR1 BULK COLLECT INTO V_RESULT;
-- collect he result into the new variable
FOR I IN 1 ..v_result.count
LOOP
dbms_output.put_line(v_result(i).<<column_name from temp offc.table>>);
end loop;
-- loop through the variable(table of records) and display its content.
-- you need to replace the << ... >> with the name of your column from source tabel that you want to display.
end;
To execute the procedure, you should use:
set serveroutput on;
execute temp_sel( 'ash');
Best,
Mikcutu
I need this piece of code in a stored procedure so I can pass an array of ids and update the related records. I am wondering whether I have to use the loop rather than use an IN clause in the sp.
SET SERVEROUTPUT ON
DECLARE
P_IDS PKGINFO.t_ids; --type: table of NUMBER index by pls_integer;
P_RESULT NUMBER;
BEGIN
p_IDS(1) := 12345;
--this works fine:
for i in ( select * from table(p_ids))
loop
UPDATE TABLE1
SET FD1 = 'test'
WHERE P_ID = i.column_value;
end loop;
--this works fine too:
SELECT COUNT(*) INTO p_RESULT FROM TABLE1
WHERE P_ID IN (SELECT * FROM TABLE (p_ids));
--but this does not work, why????? how to make it work?
UPDATE TABLE1
SET FD1 = 'test'
WHERE P_ID IN (SELECT * FROM TABLE (p_ids));
END;
--==================PKGINFO.t_ids==================
CREATE OR REPLACE package dbname.PKGINFO as
-- package created to perform Associative array calls
type t_ids is table of NUMBER index by pls_integer;
end PKGINFO;
/
I expected the UPDATE can use IN clause, but it gives me an INVALID TYPE error.
Until fairly recently Oracle didn't allow PL/SQL types in SQL statements, including in a table collection expression. You seem to be using a version where support has been added for select but not (yet) for update. If you had one available then you could use a schema-level type. Also have a look at member of.
With the type you have now you could use FORALL, which would be more efficient than a loop with individual updates::
FORALL i IN p_ids.first..p_ids.last
UPDATE TABLE1
SET FD1 = 'test'
WHERE P_ID = p_ids(i);
I have this simple query:
SELECT MEASURE_ID, MEASURE_VALUE FROM MY_TABLE;
At the moment returning just a couple of records (in the future there will plenty of them):
8 265.7
7 559.6
A DESC on such table provides:
Name Null Type
------------ -------- ------------
MEASURE_ID NOT NULL NUMBER
MEASURE_VALUE NUMBER(10,1)
Then I defined the proper PL/SQL types:
CREATE OR REPLACE TYPE HASHMAP_NUM_TYPE_OBJ AS OBJECT (
THE_ID NUMBER,
THE_VALUE NUMBER(10,1)
);
CREATE OR REPLACE TYPE HASHMAP_NUM_TYPE IS TABLE OF HASHMAP_NUM_TYPE_OBJ;
And tried to fetch the records using a BULK COLLECT:
stats_by_measure HASHMAP_NUM_TYPE;
...
OPEN cursor_1 FOR
SELECT MEASURE_ID, MEASURE_VALUE
FROM MY_TABLE;
...
FETCH cursor_1 BULK COLLECT INTO stats_by_measure;
...
CLOSE cursor_1;
But I have the Oracle -6504 error. What am I doing wrong?
Remark: If I fetch the same cursor row by row, using a codeblock like this:
foo NUMBER;
faa NUMBER(10,1);
my_obj HASHMAP_NUM_TYPE_OBJ;
...
LOOP
FETCH cursor_1 INTO foo, faa;
my_obj := HASHMAP_NUM_TYPE_OBJ(foo,faa);
EXIT WHEN cursor_1%NOTFOUND;
END LOOP;
everything works fine!
modify your cursor query like below so that it will have the same type
OPEN cursor_1 FOR
SELECT HASHMAP_NUM_TYPE_OBJ(MEASURE_ID, MEASURE_VALUE)
FROM MY_TABLE;
You can only BULK COLLECT objects into a table of objects. In your case:
SQL> CREATE OR REPLACE TYPE hashmap_num_type_obj AS OBJECT (
2 the_id NUMBER,
3 the_value NUMBER(10,1)
4 );
5 /
Type created
SQL> CREATE OR REPLACE TYPE hashmap_num_type IS TABLE OF hashmap_num_type_obj;
2 /
Type created
SQL> DECLARE
2 l_tab hashmap_num_type;
3 BEGIN
4 SELECT hashmap_num_type_obj(measure_id, measure_value)
5 BULK COLLECT INTO l_tab
6 FROM my_table;
7 END;
8 /
PL/SQL procedure successfully completed
I've solve your question
declare
type REC_TYPE is record (
THE_ID number,
THE_VALUE number
);
type TB_TYPE is table of REC_TYPE index by binary_integer;
TBL TB_TYPE;
cursor CURSOR_1 is
select a1.MEASURE_ID A$1, a1.MEASURE_VALUE A$2
from MY_TABLE a1;
type REF_CUR_ is ref cursor return CURSOR_1%rowtype;
CURSOR_2 REF_CUR_;
begin
open CURSOR_2 for
select a1.MEASURE_ID A$1, a1.MEASURE_VALUE A$2
from MY_TABLE a1;
fetch CURSOR_2 bulk collect into TBL ;
close CURSOR_2;
return;
end;
It's works.
I have found another way without ref cursor there (look for FETCH Statement with BULK COLLECT Clause)
You should retrieve the rows into a type based on a record type rather than an object type. The following works;
DECLARE
TYPE hashmap_num_type_rt IS RECORD
(THE_ID NUMBER,
THE_VALUE NUMBER(10,1)
);
TYPE hashmap_num_type_t IS TABLE OF hashmap_num_type_rt;
stats_by_measure hashmap_num_type_t;
BEGIN
SELECT measure_id, measure_value
BULK COLLECT INTO stats_by_measure
FROM my_table;
FOR i IN 1..stats_by_measure.COUNT
LOOP
DBMS_OUTPUT.PUT_LINE('RECORD '||TO_CHAR(i)||' : ID - '||stats_by_measure(i).the_id||' MeasureVal: '||TO_CHAR(stats_by_measure(i).the_value));
END LOOP;
END;
You could also define a cursor and the create a table type based on the type of the cursor (which is of course is still a row type rather than an object type).
If you want to use a cursor as the row type then try the following;
DECLARE
CURSOR c_measures IS
SELECT measure_id, measure_value
FROM my_table;
TYPE hashmap_num_type_t IS TABLE OF c_measures%ROWTYPE;
stats_by_measure hashmap_num_type_t;
BEGIN
OPEN c_measures;
FETCH c_measures
BULK COLLECT INTO stats_by_measure;
CLOSE c_measures;
FOR i IN 1..stats_by_measure.COUNT
LOOP
DBMS_OUTPUT.PUT_LINE('RECORD '||TO_CHAR(i)||' : ID - '||stats_by_measure(i).measure_id||' MeasureVal: '||TO_CHAR(stats_by_measure(i).measure_value));
END LOOP;
END;
I am using Toad. I have a declaration of a table in a package as follows:
TYPE MyRecordType IS RECORD
(ID MyTable.ID%TYPE
,FIELD1 MyTable.FIELD1%TYPE
,FIELD2 MyTable.FIELD2%TYPE
,FIELD3 MyTable.FIELD3%TYPE
,ANOTHERFIELD VARCHAR2(80)
);
TYPE MyTableType IS TABLE OF MyRecordType INDEX BY BINARY_INTEGER;
There is a procedure (lets say MyProcedure), that is using an object of this table type as input/output. I want to run the procedure and see the results (how the table is filled). So I am thinking I will select the results from the table:
declare
IO_table MyPackage.MyTableType;
begin
MyPackage.MyProcedure (IO_table
,parameter1
,parameter2
,parameter3);
select * from IO_table;
end;
I get the message:
Table or view does not exist (for IO_table). If I remove the select line, the procedure runs successfully, but I cannot see its results. How can I see the contents of IO_table after I call the procedure?
You cannot see the results for a PL/SQL table by using Select * from IO_table
You will need to loop through the collection in the annonymous block.
do something like, given in pseudo code below...
declare
IO_table MyPackage.MyTableType;
l_index BINARY_INTEGER;
begin
MyPackage.MyProcedure (IO_table
,parameter1
,parameter2
,parameter3);
l_index := IO_table.first;
While l_index is not null
loop
dbms_output.put_line (IO_table(l_index).id);
.
.
.
.
l_index :=IO_table.next(l_index_id);
end loop;
end;
You have to do it like this:
select * from TABLE(IO_table);
and, of course you missed the INTO or BULK COLLECT INTO clause
1) You can not use associated arrays in SELECT statement, Just nested tables or varrays declared globally.
2) You should use TABLE() expression in SELECT statement
3) You can't simply use SELECT in PL/SQL code - cursor FOR LOOP or REF CURSOR or BULK COLLECT INTO or INTO must be used.
4) The last but not least - please study the manual:
http://docs.oracle.com/cd/B28359_01/appdev.111/b28371/adobjcol.htm#ADOBJ00204
Just an example:
SQL> create type t_obj as object( id int, name varchar2(10));
2 /
SQL> create type t_obj_tab as table of t_obj;
2 /
SQL> var rc refcursor
SQL> declare
2 t_var t_obj_tab := t_obj_tab();
3 begin
4 t_var.extend(2);
5 t_var(1) := t_obj(1,'A');
6 t_var(2) := t_obj(2,'B');
7 open :rc for select * from table(t_var);
8 end;
9 /
SQL> print rc
ID NAME
---------- ----------
1 A
2 B
I have created the following object in oracle 11g.
CREATE OR REPLACE TYPE myObject as object(
fieldOne number,
fieldTwo number
);
And created a new table type of myObject;
CREATE OR REPLACE TYPE myTable IS TABLE OF myObject;
I would now like to create a new instance of myTable and add several hard-coded rows to myTable on the SQL Plus command line then pass the object to myProcedure as a parameter.
I have tried the following;
declare newTable myTable;
begin
select myObject(50,5) bulk collect into newTable from dual;
select myObject(40,7) bulk collect into newTable from dual;
myProcedure(newTable);
commit;
end;
Which sort-of works although the second select into statement overwrites the first.
My question is; how can I add multiple rows to newTable?
Many Thanks in Advance :)
declare
newTable myTable;
begin
newTable := myTable();
newTable.extend(2); -- The desired size of the collection
-- Oracle collections begin at index 1, not 0
newTable(1) := myObject(50, 5);
newTable(2) := myObject(40, 7);
myProcedure(newTable);
end;