I have a data base with daily tables. I want to write a procedure that will create a view to union all the tables for the days of the month. I want to catch months with 31 days or 30, and FEB with 29 days. Is it efficient to use if statement ?
Can I use if statement with more than one value as below?
if(upper(MON)) in ('JAN','MAR','MAY','JUL','AUG','OCT','DEC') THEN {DO SOMETHING}
Given how convoluted your code is, the efficiency of the IF is probably the least of your problems. I guess your company won't spring for the Partitioning licence. (Author's note: this sentence refers to the code posted in the original version of the query which the OP has since redacted.)
However, your approach won't work. You can't use an IF in DDL like that. You need to do something else.
Your view won't compile unless all the tables already exist. So I'm going to assume you have some form of scheduled job which creates those tables and then creates the view. In which case, you can drive the procedure off the data dictionary (Caution: untested code, you may need to fix typos).
CREATE OR REPLACE PROCEDURE monthly_view (
MON_P VARCHAR2,
YEAR_P VARCHAR2 DEFAULT TO_CHAR (SYSDATE, 'YY'))
IS
l_stmt varchar2(32767);
BEGIN
l_stmt := 'CREATE OR REPLACE VIEW monthly_view_'|| UPPER (MON_P)|| '_'|| YEAR_P||' as ';
for r in ( select table_name from user_tables
where table_name like 'LT_'|| UPPER (MON)||'____'|| YEAR_P||'_BCH'
)
loop
if r.table_name != 'LT_'|| UPPER (MON)||'_01_'|| YEAR_P||'_BCH'
then
l_stmt := l_stmt ||' UNION ALL ';
end if;
l_stmt := l_stmt ||' SELECT * FROM '||r.table_name;
end loop;
execute immediate l_stmt;
END;
/
This approach means only the routine which creates the tables needs to worry about the days of the month and whether it's a leap year.
Even better would be to not have separate tables for each month of each year.
Related
I am working on an application and made the decision that all the queries would be procedures. I hope to have gains in performance and ease of maintenance by doing it this way. Our DBA's have also expressed interest in having it done this way.
I have an HR table where operations are performed on it each night, and any changes are recorded in a secondary table. We are not doing auditing, these change records are kept until the next run and show users the changes that have happened.
To keep my question shorter I have reduced the number of columns in HR.
The HR table ID, GROUP_NAME, and GROUP_LEVEL. The Drill table has ID and TYPEVALUE.
CREATE OR REPLACE PROCEDURE DOCSADM.DRILL_RECORD_POSITION (
RECORD_TYPE IN VARCHAR2,
OUT_ID OUT VARCHAR2,
OUT_GROUP_NAME OUT VARCHAR2,
OUT_GROUP_LEVEL OUT VARCHAR2
) AS
BEGIN
SELECT HR.ID, HR.GROUP_NAME, HR.GROUP_LEVEL
INTO OUT_ID, OUT_GROUP_NAME, OUT_GROUP_LEVEL
FROM HR_POSITION HR JOIN DRILL_POSITION DP ON (HR.ID = DP.ID) WHERE DP.TYPEVALUE = RECORD_TYPE;
END DRILL_RECORD_POSITION;
The procedure compiles without issue. Before doing all the work in the application to link to the procedure and extract the values which in this case will eventually be displayed in a view or webpage, I wanted to have a quick little script that would call the procedure and then display the results so I can verify in Oracle.
Loops
BEGIN
for t in (DRILL_RECORD_POSITION('D', V1,V5,V6))
loop
--dbms_output.put_line(t.V1 || t.V5 || t.V6);
dbms_output.put_line(t.OUT_ID);
end loop;
END;
/
CURSORS
DECLARE
V1 HR_POSITION.ID%TYPE;
V5 HR_POSITION.GROUP_NAME%TYPE;
V6 HR_POSITION.GROUP_LEVEL%TYPE;
CURSOR T_CUR IS DRILL_RECORD_POSITION('D', V1,V5,V6);
BEGIN
OPEN T_CUR;
DBMS_OUTPUT.PUTLINE('START');
LOOP
FETCH T_CUR INTO V1,V5,V6;
EXIT WHEN T_CUR%NOTFOUND;
DBMS_OUTPUT.PUTLINE(V1||V5||V6);
END LOOP;
CLOSE T_CUR;
END;
FOR LOOPS
DECLARE
V1 HR_POSITION.POSITION_ID%TYPE;
V5 HR_POSITION.GROUP_NAME%TYPE;
V6 HR_POSITION.GROUP_LEVEL%TYPE;
BEGIN
DBMS_OUTPUT.PUTLINE('START');
FOR INDEX IN (DRILL_RECORD_POSITION('D', V1,V5,V6))
LOOP
--DBMS_OUTPUT.PUTLINE(INDEX.ID);
DBMS_OUTPUT.PUTLINE(INDEX.V1||INDEX.V5||INDEX.V6);
END LOOP;
END;
Note: I edited the column names out and shorted some when transferring here so I might have made a few mistakes.
All the articles I have seen online show me how to display from within the original procedure or by using views, cursors, records. Unless I am wrong, Eclipse wont have any problems using the information in the current form which is why I am passing it that way. So I am not interested in changing the procedure and would like to work with it as is, since thats how the application will be doing it.
As this is the first of the stored procedures I am doing for the application, instead of using adhoc queries from the application, I dont have any existing examples to work from, which is why I believe the results will work fine, because it should be the same format the adhoc ones use.
Update:
In one of the comments, I was pointed to what should have been a solution. This was confirmed by another solution that was under it.
I keep getting the error
ORA-01422: exact fetch returns more than requested number of rows
So Im returning multiple rows, but that is my expectation and what is happening. I just cant seem to figure out how to display the results.
To test the procedure you showed, you would do something like:
declare
l_id hr_position.id%type;
l_group_name hr_position.group_name%type;
l_group_level hr_position.group_level%type;
begin
drill_record_position('D', l_id, l_group_name, l_group_level);
dbms_output.put_line(l_id ||':'|| l_group_name ||':'|| l_group_level);
end;
/
But that - or more specifically, your procedure - only works if there is exactly one row in the query's result set for the passed-in value type. It seems you're expecting multiple rows back (which would get too-many-rows), but there could also be non (which would get no-data-found).
So really it seems like your question should be about how to write your procedure so it works with one of the retrieval/test methods you tried.
If your procedure needs to return multiple rows then it can use a ref cursor, e.g.:
create or replace procedure drill_record_position (
p_record_type in varchar2,
p_ref_cursor out sys_refcursor
)
as
begin
open p_ref_cursor for
select hr.id, hr.group_name, hr.group_level
from hr_position hr
join drill_position dp
on hr.id = dp.id
where dp.typevalue = p_record_type;
end drill_record_position;
/
which you could then test with something like:
declare
l_ref_cursor sys_refcursor;
l_id hr_position.id%type;
l_group_name hr_position.group_name%type;
l_group_level hr_position.group_level%type;
begin
drill_record_position('D', l_ref_cursor);
loop
fetch l_ref_cursor into l_id, l_group_name, l_group_level;
exit when l_ref_cursor%notfound;
dbms_output.put_line(l_id ||':'|| l_group_name ||':'|| l_group_level);
end loop;
close l_ref_cursor;
end;
/
You can also do that as a function, which might be easier to work with from your application:
-- drop procedure drill_record_position;
create or replace function drill_record_position (p_record_type in varchar2)
return sys_refcursor as
l_ref_cursor sys_refcursor;
begin
open l_ref_cursor for
select hr.id, hr.group_name, hr.group_level
from hr_position hr
join drill_position dp
on hr.id = dp.id
where dp.typevalue = p_record_type;
return l_ref_cursor;
end drill_record_position;
/
declare
l_ref_cursor sys_refcursor;
l_id hr_position.id%type;
l_group_name hr_position.group_name%type;
l_group_level hr_position.group_level%type;
begin
l_ref_cursor := drill_record_position('D');
loop
fetch l_ref_cursor into l_id, l_group_name, l_group_level;
exit when l_ref_cursor%notfound;
dbms_output.put_line(l_id ||':'|| l_group_name ||':'|| l_group_level);
end loop;
close l_ref_cursor;
end;
/
You coudl also do this with collections and a pipelined function, which is more work to set up:
create type t_drill_obj as object (
-- use your real data types...
id number,
group_name varchar2(10),
group_level number
)
/
create type t_drill_tab as table of t_drill_obj
/
create or replace function drill_record_position (p_record_type in varchar2)
return t_drill_tab pipelined as
begin
for l_row in (
select t_drill_obj(hr.id, hr.group_name, hr.group_level) as obj
from hr_position hr
join drill_position dp
on hr.id = dp.id
where dp.typevalue = p_record_type
)
loop
pipe row (l_row.obj);
end loop;
return;
end drill_record_position;
/
but you could call it as part of another query, and even join tot he result if you needed to:
select * from table(drill_record_position('D'));
I want to create procedure, that will use cursor, which is the same for arbitrary tables. My current one looks like this:
create or replace
PROCEDURE
some_name(
p_talbe_name IN VARCHAR2,
p_chunk_size IN NUMBER,
p_row_limit IN NUMBER
) AS
CURSOR v_cur IS
SELECT common_column,
ora_hash(substr(common_column, 1, 15), p_chunk_size) as chunk_number
-- Here it can find the table!
FROM p_table_name;
TYPE t_sample IS TALBE OF v_cur%rowtype;
v_sample t_sample;
BEGIN
OPEN v_cur;
LOOP FETCH v_cur BULK COLLECT INTO v_sample LIMIT p_row_limit;
FORALL i IN v_sample.first .. v_sample.last
INSERT INTO chunks VALUES v_sample(i);
COMMIT;
EXIT WHEN v_cur%notfound;
END LOOP;
CLOSE v_cur;
END;
The problem is that it cannot find the table named p_table_name which I want to parametrize. The thing is that I need to create chunks based on hashes for common_column which exists in all intended tables. How to deal with that problem? Maybe there is the equivalent oracle code that will do the same thing? Then I need the same efficiency for the query. Thanks!
I would do this as a single insert-as-select statement, complicated only by the fact you're passing in the table_name, so we need to use dynamic sql.
I would do it something like:
CREATE OR REPLACE PROCEDURE some_name(p_table_name IN VARCHAR2,
p_chunk_size IN NUMBER,
p_row_limit IN NUMBER) AS
v_table_name VARCHAR2(32); -- 30 characters for the tablename, 2 for doublequotes in case of case sensitive names, e.g. "table_name"
v_insert_sql CLOB;
BEGIN
-- Sanitise the passed in table_name, to ensure it meets the rules for being an identifier name. This is to avoid SQL injection in the dynamic SQL
-- statement we'll be using later.
v_table_name := DBMS_ASSERT.ENQUOTE_LITERAL(p_table_name);
v_insert_sql := 'insert into chunks (common_column_name, chunk_number)'||CHR(10)|| -- replace the column names with the actual names of your chunks table columns.
'select common_column,'||CHR(10)||
' ora_hash(substr(common_column, 1, 15), :p_chunk_size) AS chunk_number'||CHR(10)||
'from '||v_table_name||CHR(10)||
'where rownum <= :p_row_limit';
-- Used for debug purposes, so you can see the definition of the statement that's going to be run.
-- Remove before putting the code in production / convert to proper logging code:
dbms_output.put_line(v_insert_sql);
-- Now run the statement:
EXECUTE IMMEDIATE v_insert_sql USING p_chunk_size, p_row_limit;
-- I've included the p_row_limit in the above statement, since I'm not sure if your original code loops through all the rows once it processes the
-- first p_row_limit rows. If you need to insert all rows from the p_table_name into the chunks table, remove the predicate from the insert sql and the extra bind variable passed into the execute immediate.
END some_name;
/
By using a single insert-as-select statement, you are using the most efficient way of doing the work. Doing the bulk collect (which you were using) would use up memory (storing the data in the array) and cause extra context switches between the PL/SQL and SQL engines that the insert-as-select statement avoids.
I've come up with the following trigger to extract all the column names which are updated when a table row update statement is executed...
but the problem is if there are more columns(atleast 100 cols), the performance/efficiency comes into concern
sample trigger code:
set define off;
create or replace TRIGGER TEST_TRIGG
AFTER UPDATE ON A_AAA
FOR EACH ROW
DECLARE
mytable varchar2(32) := 'A_AAA';
mycolumn varchar2(32);
updatedcols varchar2(3000);
cursor s1 (mytable varchar2) is
select column_name from user_tab_columns where table_name = mytable;
begin
open s1 (mytable);
loop
fetch s1 into mycolumn;
exit when s1%NOTFOUND;
IF UPDATING( mycolumn ) THEN
updatedcols := updatedcols || ',' || mycolumn;
END IF;
end loop;
close s1;
--do a few things with the list of updated columns
dbms_output.put_line('updated cols ' || updatedcols);
end;
/
Is there any alternative way to get the list?
Maybe with v$ tables (v$transaction or anything similar)?
No its the best way to get UPDATED column by UPDATING()
and you can change your code using implicit cursor like this, it will be a little bit faster
set define off;
create or replace TRIGGER TEST_TRIGG
AFTER UPDATE ON A_AAA
FOR EACH ROW
DECLARE
updatedcols varchar2(3000);
begin
for r in (select column_name from user_tab_columns where table_name ='A_AAA')
loop
IF UPDATING(r.column_name) THEN
updatedcols := updatedcols || ',' || r.column_name;
END IF;
end loop;
dbms_output.put_line('updated cols ' || updatedcols);
end;
/
Faced with a similar task, we ended up writing a pl/sql procedure which lists the columns of the table and generates the full trigger body for us, with static code referencing :new.col and :old.col. The execution of such trigger should probably be faster (though we didn't compare).
However, the downside is that when you later add a new column to the table, it's easy to forget to update the trigger body. It probably can be managed somehow with a monitoring job or elsehow, but for now it works for us.
P.S. I became curious what that updating('COL') feature does, and checked it now. I found out that it returns true if the column is present in the update statement, even if the value of the column actually didn't change (:old.col is equal to :new:col). This might generate unneeded history records, if the table is being updated by something like Java Hibernate library, which (by default) always specifies all columns in the update statements it generates. In such a case you might want to actually compare the values from inside the trigger body and insert the history record only in case the new value differs from the old value.
I'm trying to use an array of input values to my procedure in an IN Clause as part of the where clause of a cursor. I know that this has been asked before, but I haven't seen how to make my syntax compile correctly.
In the package specification, the type is
TYPE t_brth_dt IS TABLE OF sourceTable.stdt_brth_dt%TYPE INDEX BY PLS_INTEGER;
sourceTable.std_brth_dt is a date column in the table.
Simplified version of my cursor is in the package body is -
cursor DataCursor_Sort( p_brth_dt in t_brth_dt) is
SELECT *
FROM sourceTable
WHERE a.brth_dt IN (select column_value
from table(p_brth_dt))
When I try to compile this, I'm getting the following errors.
[1]:(Error): PLS-00382: expression is of wrong type
[2]:(Error): PL/SQL: ORA-22905: cannot access rows from a non-nested table item
I know this looks similar to other questions, but I don't understand what the syntax error is.
In order to use collection defined as a nested table or an associative array in the from clause of a query you either should, as #Alex Poole correctly pointed out, create a schema level (SQL) type or use one, that is available to you trough ODCIConst package - odcidatelist as you intend to use a list of dates. For example, your cursor definition might look like this:
cursor DataCursor_Sort(p_brth_dt in sys.odcidatelist) is
select *
from sourceTable
where a.brth_dt IN (select column_value
from table(p_brth_dt))
OR
cursor DataCursor_Sort(p_brth_dt in sys.odcidatelist) is
select s.*
from sourceTable s
join table(p_brth_dt) t
on (s.brth_dt = t.column_value)
Note: You should take into consideration the time part of a date when performing a date comparison. If you want to compare date part only it probably would be useful to get rid of time part by using trunc() function.
It is possible to use a PL/SQL-defined nested table type (as opposed to a SQL-defined nested table type) indirectly in an IN clause of a SELECT statement in a PL/SQL package. You must use a PIPELINED function as an intermediary. It felt kind of clever to write, but I don't believe in its fundamental usefulness.
CREATE OR REPLACE PACKAGE so18989249 IS
TYPE date_plsql_nested_table_type IS TABLE OF DATE;
dates date_plsql_nested_table_type;
FUNCTION dates_pipelined RETURN date_plsql_nested_table_type PIPELINED;
PROCEDURE use_plsql_nested_table_type;
END so18989249;
/
CREATE OR REPLACE PACKAGE BODY so18989249 IS
FUNCTION dates_pipelined RETURN date_plsql_nested_table_type
PIPELINED IS
BEGIN
IF (dates.count > 0)
THEN
FOR i IN dates.first .. dates.last
LOOP
IF (dates.exists(i))
THEN
PIPE ROW(dates(i));
END IF;
END LOOP;
END IF;
END;
PROCEDURE use_plsql_nested_table_type IS
BEGIN
dates := NEW date_plsql_nested_table_type();
-- tweak these values as you see fit to produce the dbms_output results you want
dates.extend(5);
dates(1) := DATE '2013-12-25';
dates(2) := DATE '2013-01-01';
dates(3) := DATE '2013-07-01';
dates(4) := DATE '2013-09-03';
dates(5) := DATE '2008-11-18';
FOR i IN (SELECT o.owner,
o.object_name,
o.object_type,
to_char(o.last_ddl_time, 'YYYY-MM-DD') AS last_ddl
FROM all_objects o
WHERE trunc(o.last_ddl_time) IN
(SELECT column_value FROM TABLE(dates_pipelined))
--uses pipeline function which uses pl/sql-defined nested table
)
LOOP
dbms_output.put_line('"' || i.owner || '"."' || i.object_name || '" ("' || i.object_type || ') on ' || i.last_ddl);
END LOOP;
END;
END so18989249;
/
begin so18989249.use_plsql_nested_table_type; end;
/
The type has to be created at SQL level, not in a package. An SQL query doesn't know how to use any types defined in PL/SQL. So you'd have to do:
CREATE OR REPLACE TYPE t_brth_dt IS TABLE OF date;
/
... and remove the type from your package specification. (Or give them different names, at least, and they won't be interchangeable in use). Because it's at SQL level, you also can't use sourceTable.stdt_brth_dt%TYPE in the declaration, unfortunately.
For debugging purposes, I'd like to be able to "TO_CHAR" an Oracle PL/SQL in-memory table. Here's a simplified example, of what I'd like to do:
DECLARE
TYPE T IS TABLE OF MY_TABLE%ROWTYPE INDEX BY PLS_INTEGER;
V T;
BEGIN
-- ..
-- Here, I'd like to dbms_output V's contents, which of course doesn't compile
FOR i IN V.FIRST .. V.LAST LOOP
dbms_output.put_line(V(i));
END LOOP;
-- I want to omit doing this:
FOR i IN V.FIRST .. V.LAST LOOP
dbms_output.put_line(V(i).ID || ',' || V(i).AMOUNT ...);
END LOOP;
END;
Can this be achieved, simply? The reason I ask is because I'm too lazy to write this debugging code again and again, and I'd like to use it with any table type.
ok, sorry this isn't complete, but to followup with #Lukas, here's what I have so far:
First, instead of trying to create anydata/anytype types, I tried using XML extracted from a cursor...weird, but its generic:
CREATE OR REPLACE procedure printCur(in_cursor IN sys_refcursor) IS
begin
FOR c IN (SELECT ROWNUM rn,
t2.COLUMN_VALUE.getrootelement () NAME,
EXTRACTVALUE (t2.COLUMN_VALUE, 'node()') VALUE
FROM TABLE (XMLSEQUENCE (in_cursor)) t,
TABLE (XMLSEQUENCE (EXTRACT (COLUMN_VALUE, '/ROW/node()'))) t2
order by 1)
LOOP
DBMS_OUTPUT.put_line (c.NAME || ': ' || c.VALUE);
END LOOP;
exception
when others then raise;
end;
/
Now, to call it, you need a cursor, so I tried casting to cursor in pl/sql, something like:
open v_cur for select * from table(cast(v_tab as tab_type));
But depending on how v_tab is defined, this may or may not cause issues in pl/sql cast (using %rowtype in nested table def seems to give issues).
Anyway, you can build on this or refine it as you like. (and possibly use xmltable...)
Hope that helps