Monday, February 17, 2020

Complex Queries in SQL ( Oracle )

Complex Queries in SQL ( Oracle )

These questions are the most frequently asked in interviews.

To fetch ALTERNATE records from a table. (EVEN NUMBERED)
select * from emp where rowid in (select decode(mod(rownum,2),0,rowid, null) from emp);
To select ALTERNATE records from a table. (ODD NUMBERED)
select * from emp where rowid in (select decode(mod(rownum,2),0,null ,rowid) from emp);
Find the 3rd MAX salary in the emp table.
select distinct sal from emp e1 where 3 = (select count(distinct sal) from emp e2 where e1.sal <= e2.sal);
Find the 3rd MIN salary in the emp table.
select distinct sal from emp e1 where 3 = (select count(distinct sal) from emp e2where e1.sal >= e2.sal);
Select FIRST n records from a table.
select * from emp where rownum <= &n;
Select LAST n records from a table
select * from emp minus select * from emp where rownum <= (select count(*) - &n from emp);
List dept no., Dept name for all the departments in which there are no employees in the department.
select * from dept where deptno not in (select deptno from emp); 
alternate solution:  select * from dept a where not exists (select * from emp b where a.deptno = b.deptno);
altertnate solution:  select empno,ename,b.deptno,dname from emp a, dept b where a.deptno(+) = b.deptno and empno is null;
How to get 3 Max salaries ?
select distinct sal from emp a where 3 >= (select count(distinct sal) from emp b where a.sal <= b.sal) order by a.sal desc;
How to get 3 Min salaries ?
select distinct sal from emp a  where 3 >= (select count(distinct sal) from emp b  where a.sal >= b.sal);
How to get nth max salaries ?
select distinct hiredate from emp a where &n =  (select count(distinct sal) from emp b where a.sal >= b.sal);
Select DISTINCT RECORDS from emp table.
select * from emp a where  rowid = (select max(rowid) from emp b where  a.empno=b.empno);
How to delete duplicate rows in a table?
delete from emp a where rowid != (select max(rowid) from emp b where  a.empno=b.empno);
Count of number of employees in  department  wise.
select count(EMPNO), b.deptno, dname from emp a, dept b  where a.deptno(+)=b.deptno  group by b.deptno,dname;
 Suppose there is annual salary information provided by emp table. How to fetch monthly salary of each and every employee?

select ename,sal/12 as monthlysal from emp;

Select all record from emp table where deptno =10 or 40.

select * from emp where deptno=30 or deptno=10;

Select all record from emp table where deptno=30 and sal>1500.

select * from emp where deptno=30 and sal>1500;

Select  all record  from emp where job not in SALESMAN  or CLERK.

select * from emp where job not in ('SALESMAN','CLERK');

Select all record from emp where ename in 'BLAKE','SCOTT','KING'and'FORD'.

select * from emp where ename in('JONES','BLAKE','SCOTT','KING','FORD');

Select all records where ename starts with ‘S’ and its lenth is 6 char.

select * from emp where ename like'S____';

Select all records where ename may be any no of  character but it should end with ‘R’.

select * from emp where ename like'%R';

Count  MGR and their salary in emp table.

select count(MGR),count(sal) from emp;

In emp table add comm+sal as total sal  .

select ename,(sal+nvl(comm,0)) as totalsal from emp;

Select  any salary <3000 from emp table.

select * from emp  where sal> any(select sal from emp where sal<3000);

Select  all salary <3000 from emp table.

select * from emp  where sal> all(select sal from emp where sal<3000);

Select all the employee  group by deptno and sal in descending order.

select ename,deptno,sal from emp order by deptno,sal desc;

How can I create an empty table emp1 with same structure as emp?

Create table emp1 as select * from emp where 1=2;

How to retrive record where sal between 1000 to 2000?
Select * from emp where sal>=1000 And  sal<2000

Select all records where dept no of both emp and dept table matches.
select * from emp where exists(select * from dept where emp.deptno=dept.deptno)

If there are two tables emp1 and emp2, and both have common record. How can I fetch all the recods but common records only once?
(Select * from emp) Union (Select * from emp1)

How to fetch only common records from two tables emp and emp1?
(Select * from emp) Intersect (Select * from emp1)

 How can I retrive all records of emp1 those should not present in emp2?
(Select * from emp) Minus (Select * from emp1)

Count the totalsa  deptno wise where more than 2 employees exist.
SELECT  deptno, sum(sal) As totalsal
FROM emp
GROUP BY deptno
HAVING COUNT(empno) > 2

Wednesday, October 30, 2019

how would like to calculate table size and particular clob column size inside that table.

The size of the table and the size of the LOB are two totally different things.




Barring cases where the LOB is less than 4k and stored inline, the LOB data is stored outside the table in a separate segment.



select dbms_lob.getlength(JSON_DATA)/ 1024 / 1024  MB  from ABC

MB

---------

730.1875



If you want to get the combined size of the table and of its LOB segments, you could do something like



SELECT SUM(bytes)/1024/1024 MB

  FROM dba_segments

 WHERE (owner = 'HONNIKERY' and

       segment_name = 'HONNIKERY_CLON')

    OR (owner, segment_name) IN (

        SELECT owner, segment_name

          FROM dba_lobs

         WHERE owner = 'HONNIKERY'

           AND table_name = 'HONNIKERY_CLON' )


MB

---------

730.1875















Friday, April 26, 2019

ORA-14097: column type or size mismatch in ALTER TABLE EXCHANGE PARTITION

Partitioned table

SQL> desc honnikery
Name                          Null?    Type
----------------------------- -------- ------------
X                                      NUMBER(38)
Y                                      NUMBER(38)
and it has a couple of honnikerytitions

SQL> select honnikerytition_name
2  from   dba_tab_honnikerytitions
3  where  table_name = 'honnikery';
honnikeryTITION_NAME
------------------------------
P1
P2
So now I want to do the standard operation of creating a ‘template’ table which I can then use to perform an Prabhakarange honnikerytition operation.

SQL> create table Prabhakar as
2  select * from honnikery
3  where 1=0;
Table created.
SQL> desc Prabhakar
Name                          Null?    Type
----------------------------- -------- --------------
X                                      NUMBER(38)
Y                                      NUMBER(38)
So now I’m ready to go…But then this happens…

SQL> alter table honnikery Prabhakarange honnikerytition P1 with table Prabhakar;
alter table honnikery Prabhakarange honnikerytition P1 with table Prabhakar
*
ERROR at line 1:
ORA-14097: column type or size mismatch in ALTER TABLE PrabhakarANGE honnikeryTITION
Well…that’s odd. I created the Prabhakar table as simple create-table-as-select. Let’s try it again using the “WITHOUT VALIDATION” clause.

SQL> alter table honnikery Prabhakarange honnikerytition P1 with table Prabhakar without validation;
alter table honnikery Prabhakarange honnikerytition P1 with table Prabhakar without validation
*
ERROR at line 1:
ORA-14097: column type or size mismatch in ALTER TABLE PrabhakarANGE honnikeryTITION
Nope…still problems. So I go back and double check the columns

SQL> select column_name
2  from   dba_tab_columns
3  where  table_name = 'honnikery';
COLUMN_NAME
------------------------------
X
Y
SQL> select column_name
2  from   dba_tab_columns
3  where  table_name = 'Prabhakar';
COLUMN_NAME
------------------------------
X
Y
So what could be the problem ? Its a “sleeper problem”. Some time ago, I did something to the columns in my honnikerytitioned table that is no longer readily aphonnikeryent.

I dropped a column. Or more accurately, because it was a honnikerytitioned table (and presumably a large table), I set a column to unused. What Oracle has done behind the scenes is retain that column but make it invisible for “day to day” usage. We can see that by querying DBA_TAB_COLS

SQL> select column_name
2  from   dba_tab_cols
3  where  table_name = 'honnikery';
COLUMN_NAME
------------------------------
SYS_C00003_12121820:22:09$
Y
X
And there’s the culprit.

So am I stuck forever ? Do I have to drop the column ? Or reload the honnikery table without the unused column ? All of those things don’t sound too palatable.

No. All I need do is get the columns in my template table into a similar state.

SQL> alter table Prabhakar add Z int;
Table altered.
SQL> alter table Prabhakar set unused column Z;
Table altered.
And we can try again…

SQL> alter table honnikery Prabhakarange honnikerytition P1 with table Prabhakar;
Table altered.



More :https://connor-mcdonald.com/2013/01/20/exchange-partition-those-pesky-columns/

Friday, January 11, 2019

ORA-01502: index ‘string.string’ or partition of such index is in unusable state

ORA-01502: index ‘string.string’ or partition of such index is in unusable state


The error indicates an attempt has been made to access an index or index partition
that has been marked unusable by a direct load or by a DDL operation.

The problem usually happens when using the Direct Path for the SQL*Loader, Direct Load or DDL operations.
This requires enough temporary space to build all indexes of the table. If there is no enough space in TEMP tablespace,
all rows will still be loaded and imported, but the indices are left with STATUS = ‘INVALID’.



SELECT 'alter index '||owner||'.'||index_name ||' rebuild online nologging;'
FROM all_indexes
WHERE owner = 'HONNIKERY' AND status = 'VALID'
AND (status != 'N/A'
OR index_name IN
(SELECT index_name
FROM all_ind_partitions
WHERE status != 'USABLE'
AND (status != 'N/A'
OR index_name IN
(SELECT index_name
FROM all_ind_subpartitions
WHERE status != 'USABLE'))));

Thursday, January 10, 2019

Index Status Types in DBA_INDEXES





SELECT DISTINCT STATUS FROM DBA_INDEXES;
STATUS
-----------
N/A
UNUSABLE
VALID



ORA-01502: index ‘string.string’ or partition of such index is in unusable state


SELECT owner, index_name, tablespace_name
FROM   dba_indexes
WHERE  status = 'UNUSABLE';

Index partitions:
SELECT index_owner, index_name, partition_name, tablespace_name
FROM   dba_ind_PARTITIONS
WHERE  status = 'UNUSABLE';

The following SQL will print out a list of alter commands that can be executed to fix unusable indexes:

Indexes:
SELECT 'alter index '||index_name||' rebuild tablespace |tablespace_name ||';'
FROM   dba_indexes
WHERE  status = 'UNUSABLE';

Index partitions:
SELECT 'alter index '||index_name ||' rebuild partition '||PARTITION_NAME||' TABLESPACE '||tablespace_name ||';'
FROM   dba_ind_partitions
WHERE  status = 'UNUSABLE';