Example scenario:
5 tables are there and one common field among them is com_field (DATE data type). Now, I need to find the maximum of com_field in each of the five tables. Can someone give the logic?
I know UNION could be used but I need the flexibility not to miss any new table added to the OWNER.
The result I am expecting is like the below.
Table Max(com_field)
Tbl1 10/21/2019
Tbl2 10/18/2019
Tbl3 10/28/2019
Tbl4 09/30/2019
Tbl5 09/09/2019
Run this query:
SELECT
'SELECT '''||OWNER||'.'||TABLE_NAME ||''' AS TABLE_NAME , '||'MAX(COM_FIELD)AS COM_FIELD FROM ' ||OWNER||'.'||TABLE_NAME ||' UNION ALL'
FROM ALL_TAB_COLUMNS
WHERE COLUMN_NAME ='COM_FIELD'
then copy the outpu and delete last union all keyword. Then run the sql statement.
You can order it and see the max value
One option is to use dynamic SQL in a function that returns refcursor. Here's an example.
First, test case:
SQL> create table taba (com_field date);
Table created.
SQL> create table tabb (com_field date);
Table created.
SQL> create table tabc (com_field date);
Table created.
SQL> insert all
2 into taba values (sysdate)
3 into taba values (sysdate - 2)
4 into taba values (sysdate - 3)
5 into tabb values (sysdate + 2)
6 into tabc values (sysdate + 4)
7 into tabc values (sysdate + 5)
8 select * From dual;
6 rows created.
SQL>
Function:
SQL> create or replace function f_maxcom
2 return sys_refcursor
3 is
4 l_str varchar2(1000);
5 rc sys_refcursor;
6 begin
7 for cur_r in (select table_name
8 from user_tab_columns
9 where column_name = 'COM_FIELD'
10 )
11 loop
12 l_str := l_str ||
13 'select ' || chr(39) || cur_r.table_name || chr(39) || ', ' ||
14 'max(com_field) from ' || cur_r.table_name || ' union all ';
15 end loop;
16
17 l_str := rtrim(l_str, ' union all');
18
19 open rc for l_str;
20 return rc;
21 end;
22 /
Function created.
SQL>
Let's try it:
SQL> select f_maxcom from dual;
F_MAXCOM
--------------------
CURSOR STATEMENT : 1
CURSOR STATEMENT : 1
'TAB MAX(COM_FI
---- ----------
TABA 29.10.2019
TABB 31.10.2019
TABC 03.11.2019
SQL>
Add another table to see what happens; function will stay as is:
SQL> create table littlefoot (id number, com_field date);
Table created.
SQL> insert into littlefoot values (100, sysdate);
1 row created.
SQL> select f_maxcom from dual;
F_MAXCOM
--------------------
CURSOR STATEMENT : 1
CURSOR STATEMENT : 1
'LITTLEFOO MAX(COM_FI
---------- ----------
LITTLEFOOT 29.10.2019
TABA 29.10.2019
TABB 31.10.2019
TABC 03.11.2019
SQL>
Seems to be OK, eh?
Related
Let's say that I have a "Rules" table that has a column that contains oracle queries in a varchar2 column:
Row
Query
1
select 'Hardcoded_Tablename_1' tablename, request_id from table_1 where status >= 4 and resolve_date < *some date math goes here*
2
select 'Table_2' tablename, t2.request_id from table_2 t2 join table_1 t1 on t1.request_id = t2.parent_id where t1.status >= 4 and t1.resolve_date < *some date math goes here*
If this were never going to change, I'd just make a view with a union of these queries.
Our requirement is that we be able to add to or to modify these rules on-the-fly at the whims of leadership.
So, what I need is either:
a very smart view (I think impossible) that executes and unions all of these stored query strings
or
a table function that returns the results of the union of these stored query strings. (I think this is the more likely solution)
It will only ever be those two columns: The hardcoded name of the table and the ID of the record.
Can someone help get me started on this?
Thanks
You can use a PIPELINED function.
First create the types:
CREATE TYPE request_data IS OBJECT (tablename VARCHAR2(30), request_id NUMBER);
CREATE TYPE request_list IS TABLE OF request_data;
Then the function:
CREATE FUNCTION get_requests RETURN request_list PIPELINED
IS
BEGIN
FOR r IN (SELECT "QUERY" FROM table_name ORDER BY "ROW")
LOOP
DECLARE
c_cursor SYS_REFCURSOR;
v_tablename VARCHAR2(30);
v_request_id NUMBER;
BEGIN
OPEN c_cursor FOR r."QUERY";
LOOP
FETCH c_cursor INTO v_tablename, v_request_id;
EXIT WHEN c_cursor%NOTFOUND;
PIPE ROW (request_data(v_tablename, v_request_id));
END LOOP;
CLOSE c_cursor;
EXCEPTION
WHEN NO_DATA_NEEDED THEN
CLOSE c_cursor;
RETURN;
END;
END LOOP;
END;
/
Then, if you have the sample data:
CREATE TABLE table_name ("ROW", "QUERY") AS
SELECT 1, q'[select 'Hardcoded_Tablename_1' tablename, request_id from table_1 where status >= 4 and resolve_date < SYSDATE]' FROM DUAL UNION ALL
SELECT 2, q'[select 'Table_2' tablename, t2.request_id from table_2 t2 join table_1 t1 on t1.request_id = t2.parent_id where t1.status >= 4 and t1.resolve_date < SYSDATE]' FROM DUAL
CREATE TABLE table_1 (request_id, status, resolve_date) AS
SELECT 42, 4, SYSDATE - 1 FROM DUAL;
CREATE TABLE table_2 (request_id, parent_id) AS
SELECT 57, 42 FROM DUAL;
Then you can use the function in a table collection expression:
SELECT *
FROM TABLE(get_requests());
Which outputs:
TABLENAME
REQUEST_ID
Hardcoded_Tablename_1
42
Table_2
57
db<>fiddle here
One option might be a function that returns refcursor.
SQL> select * from rules;
CROW QUERY
---------- ----------------------------------------------------------------------------------------------------
1 select 'EMP' tablename, empno from emp where hiredate = (select max(hiredate) from emp)
2 select 'DEPT' tablename, d.deptno from emp e join dept d on d.deptno = e.deptno where e.hiredate = (
select min(hiredate) from emp)
Function creates union of all queries from the rules table and uses it as a source for the refcursor:
SQL> create or replace function f_test return sys_refcursor
2 is
3 l_rc sys_refcursor;
4 l_str clob;
5 begin
6 for cur_r in (select query from rules order by crow) loop
7 l_str := l_str || cur_r.query ||' union all ';
8 end loop;
9 l_str := rtrim(l_str, ' union all ');
10
11 open l_rc for l_str;
12 return l_rc;
13 end;
14 /
Function created.
Testing:
SQL> select f_test from dual;
F_TEST
--------------------
CURSOR STATEMENT : 1
CURSOR STATEMENT : 1
TABL EMPNO
---- ----------
EMP 7876
DEPT 20
SQL>
I have a table "TEST_TABLE" with two columns TABLE_NAME and RECORD_COUNT.
enter image description here
We want to update the column RECORD_COUNT by taking the total records in table specified in TABLE_NAME.
You could do this with dynamic SQL but why be complicated?
Create a view
Create view my_tables as
Select
'table_1' as "table name",
count(*) as "rows"
From table_1
Add the following for each table
Union all
Select 'table_2',count(*) From table_2
You can then use the view like a table:
Select * from my_tables;
OK, but - why wouldn't you use information Oracle already provides for you? You should regularly gather statistics anyway, so:
SQL> execute dbms_stats.gather_schema_stats(ownname => 'SCOTT', estimate_percent => null);
PL/SQL procedure successfully completed.
and then fetch num_rows from user_tables:
SQL> select table_name, num_rows from user_tables where rownum <= 10;
TABLE_NAME NUM_ROWS
-------------------- ----------
EMP 14
DEPT 4
BONUS 0
SALGRADE 5
DUMMY 1
TBL_ERROR 1
AUDIT_TAB 2
SOURCE_DET 3
EXAMPLE 1
FLIGHT 0
10 rows selected.
SQL>
It can't be much worse than your attempt (due to possible frequent changes to tables' contents; inserts and deletes) because you'd collect your own data periodically (how often?) anyway.
If it has to be your way, then you'd use dynamic SQL, looping through all tables in your schema and actually count number of rows:
SQL> create table test_table
2 (table_name varchar2(30),
3 num_rows number);
Table created.
SQL> create or replace procedure p_test as
2 l_cnt number;
3 begin
4 execute immediate 'truncate table test_table';
5 for cur_R in (select table_name from user_tables) loop
6 execute immediate 'select count(*) from ' ||cur_R.table_name into l_Cnt;
7 insert into test_table (table_name, num_rows) values (cur_r.table_name, l_cnt);
8 end loop;
9 end;
10 /
Procedure created.
Running the procedure and viewing the result:
SQL> exec p_test;
PL/SQL procedure successfully completed.
SQL> select * From test_Table where rownum <= 10;
TABLE_NAME NUM_ROWS
-------------------- ----------
EMP 14
DEPT 4
BONUS 0
SALGRADE 5
DUMMY 1
TBL_ERROR 1
AUDIT_TAB 2
SOURCE_DET 3
EXAMPLE 1
FLIGHT 0
10 rows selected.
SQL>
Note that performance will suffer as number of tables and number of rows stored in them grows.
If I were you, I'd go for the first option.
How to fetch odd columns in Oracle using a query when number of columns and name of columns are not known?
E.g.:
I need to get output in below format
Column1 column3 column5 column7
And so on....
You need to use the dynamic queries in the procedure as follows:
SQL> CREATE OR REPLACE PROCEDURE ODD_COLUMNS (
2 TABLE_NAME_P IN VARCHAR2,
3 DATAA OUT SYS_REFCURSOR
4 ) AS
5 V_SQL VARCHAR2(4000);
6 BEGIN
7 SELECT
8 'SELECT '
9 ||
10 LISTAGG(COLUMN_NAME, ',') WITHIN GROUP(
11 ORDER BY
12 COLUMN_ID
13 )
14 || ' FROM "'
15 || TABLE_NAME_P
16 || '"'
17 INTO V_SQL
18 FROM
19 USER_TAB_COLS
20 WHERE
21 TABLE_NAME = TABLE_NAME_P
22 AND MOD(COLUMN_ID, 2) = 1;
23
24 OPEN DATAA FOR V_SQL;
25
26 END ODD_COLUMNS;
27 /
Procedure created.
SQL>
Now, Let's test it:
SQL> variable rc refcursor;
SQL> exec ODD_COLUMNS('EMP',:rc);
PL/SQL procedure successfully completed.
SQL> print rc;
EMP_ID E
---------- -
10 N
20 Y
SQL>
SQL> exec ODD_COLUMNS('MY_TABLE1',:rc);
PL/SQL procedure successfully completed.
SQL> print rc;
ID REQ_QTY
---------- ----------
1001 10
1001 20
1001 30
1002 40
1003 10
1003 20
6 rows selected.
SQL>
Cheers!!
This cannot be done simply, but it is possible using the Oracle data dictionary and some dynamic SQL.
To find out the odd-numbered columns you need to look at the ALL_TAB_COLUMNS view. Column COLUMN_ID sequences the columns 1,2,3. So this will find all the odd-numbered columns in the SCOTT.EMP table:
select column_name, column_id
from all_tab_columns
where owner = 'SCOTT'
and table_name = 'EMP'
and mod(column_id,2) = 1
order by column_id;
This will return something like:
COLUMN_NAME COLUMN_ID
----------- ---------
EMPNO 1
JOB 3
HIREDATE 5
COMM 7
We can use the LISTAGG function to make that into a comma-separated list:
select listagg(column_name,',') within group (order by column_id) as result
from user_tab_columns
where table_name = 'EMP'
and mod(column_id,2) = 1;
RESULT
------
EMPNO,JOB,HIREDATE,COMM
Now we can add to that SQL to generate the select statement you want:
select 'select ' || listagg(column_name,',') within group (order by column_id) || ' from ' || table_name as sql
from user_tab_columns
where table_name = 'EMP'
and mod(column_id,2) = 1
group by table_name;
SQL
---
select EMPNO,JOB,HIREDATE,COMM from EMP
(Note I had to add a group by clause because table_name is not being aggregated by LISTAGG).
You could use that SQL within some PL/SQL code to populate a variable v_sql, then use the DBMS_SQL package to run it. But that is a complex topic in itself and I won't go into it here.
I want to run a query on different schemes to get data and export it. I use the following code
DECLARE
sql_statment VARCHAR2(2000);
BEGIN
FOR c IN (SELECT brchcode FROM brchs) LOOP
sql_statment := 'select distinct ''' || c.brchcode ||''', t.risuid from ' || c.brchcode ||
'.reg_individualacnt_detail t
where t.historytypecode = 60';
EXECUTE IMMEDIATE sql_statment;
END LOOP;
END;
where brchcode is the name of different schemes
I can't see any output. what can I do?
Code you wrote can't work as you have to return the result into something; it is PL/SQL and requires an INTO clause. As you chose to return two values (columns) and multiple rows, that can't be a scalar variable; you could pick ref cursor or a collection, for example.
Here's an example which shows one option.
I'll be using two schemas: SCOTT (current schema) and HR. Both will be having the DEPT table.
As Scott already has it, I'll create one in HR schema and grant access to Scott (otherwise, Scott won't even see it and the procedure (i.e. the function) will fail):
SQL> connect hr/hr
Connected.
SQL> create table dept (deptno number, dname varchar2(10), loc varchar2(10));
Table created.
SQL> insert into dept values (55, 'IT', 'Zagreb');
1 row created.
SQL> grant select on dept to scott;
Grant succeeded.
SQL> commit;
Commit complete.
Back to Scott, to create a table (which contains schema names I'll be selecting from) and a function. I chose to return REF CURSOR; you could return something else, if you want.
SQL> connect scott/tiger
Connected.
SQL> create table brchs (brchcode varchar2(10));
Table created.
SQL> insert into brchs (brchcode)
2 select 'scott' from dual union all
3 select 'hr' from dual;
2 rows created.
SQL> create or replace function f_br
2 return sys_refcursor
3 is
4 l_str varchar2(4000);
5 l_rc sys_refcursor;
6 begin
7 for cur_r in (select brchcode from brchs) loop
8 l_str := l_str ||
9 'union all
10 select ' || chr(39)|| cur_r.brchcode ||chr(39) || ', d.dname
11 from ' || cur_r.brchcode ||'.dept d
12 where d.deptno > 0';
13 end loop;
14
15 l_str := ltrim(l_str, 'union all');
16
17 open l_rc for l_str;
18 return l_rc;
19 end;
20 /
Function created.
SQL>
Finally, testing:
SQL> select f_br from dual;
F_BR
--------------------
CURSOR STATEMENT : 1
CURSOR STATEMENT : 1
'SCOT DNAME
----- --------------
scott ACCOUNTING
scott RESEARCH
scott SALES
scott OPERATIONS
hr IT
SQL>
I have one table and I need to create one dummy table with same column names and data but with different datatypes for some of the columns.
For example: Table-1 has two columns C1 (varchar2) and C2(date).
I need to create a dummy table called Table-2 with columns C1 (varchar2) and C2(varchar2).
Please suggest the way to do it in oracle.
The best way to do this is to duplicate the table with create as select, without the data, for example -
create table Table-2 as select * from Table-1 where 1=0;
And then alter the datatypes of the required columns manually like so -
alter table Table-2 modify (C2 varchar2);
After the column was altered you can push the data from Table-1 into Table-2, using proper conversions. in your example -
insert into Table-2 select C1, to_char(C2,'dd-mm-yyyy') from Table-1;
Assuming that all the columns of the starting table can be converted ( implicit conversion) in VARCHAR2, you can do something like the following.
Say you have this table:
SQL> create table table1 (
2 date_field date,
3 varchar_field varchar2(1000),
4 number_field number
5 );
Table created.
SQL> insert into table1 values (sysdate, 'some text', 999);
1 row created.
SQL> commit;
Commit complete.
You can build a dynamis SQL that creates another table and copies the data from one table to another, using implicit type conversion:
SQL> declare
2 stm varchar2(32767);
3 begin
4 select 'create table table2( ' ||
5 listagg(column_name, ' varchar2(4000), ') within group (order by column_name) ||
6 ' varchar2(4000) )'
7 into stm
8 from user_tab_columns
9 where table_name = 'TABLE1';
10 --
11 execute immediate stm;
12 --
13 select 'insert into table2( ' ||
14 listagg(column_name, ', ') within group (order by column_name) ||
15 ' ) select ' ||
16 listagg(column_name, ', ') within group (order by column_name) ||
17 ' from table1'
18 into stm
19 from user_tab_columns
20 where table_name = 'TABLE1';
21 execute immediate stm;
22 end;
23 /
PL/SQL procedure successfully completed.
SQL> select * from table2;
DATE_FIELD NUMBER_FIELD VARCHAR_FIELD
--------------- --------------- ---------------
27-APR-16 some text 999
SQL>