How to find top three salary of emp table?

Hi,
I am trying to get top 3 salaries using MAX() in sql but it is only showing 1 row not 3.The following qury showing only the 3rd highest.
SELECT MAX(Sal) ThirdHighest FROM emp
WHERE Sal < (SELECT MAX(Sal) FROM emp
WHERE Sal =(SELECT MAX(Sal) FROM emp
WHERE Sal <(SELECT MAX(Sal) FROM emp)))
I dont want to use the following method.
SELECT department_id dept, first_name || ' ' || last_name name, salary
FROM (
SELECT department_id, first_name, last_name,
MAX(salary) OVER (PARTITION BY department_id) dept_max_sal, salary
FROM employees e
) WHERE salary = dept_max_sal;
Kindly help to create query using only "max()" for top 3 salaries.

947459 wrote:
Hi,
I am trying to get top 3 salaries using MAX() in sql but it is only showing 1 row not 3.The following qury showing only the 3rd highest.
SELECT MAX(Sal) ThirdHighest FROM emp
WHERE Sal < (SELECT MAX(Sal) FROM emp
WHERE Sal =(SELECT MAX(Sal) FROM emp
WHERE Sal <(SELECT MAX(Sal) FROM emp)))
I dont want to use the following method.
SELECT department_id dept, first_name || ' ' || last_name name, salary
FROM (
SELECT department_id, first_name, last_name,
MAX(salary) OVER (PARTITION BY department_id) dept_max_sal, salary
FROM employees e
) WHERE salary = dept_max_sal;
Kindly help to create query using only "max()" for top 3 salaries.Why? Why not use what Oracle provides for doing top N queries?
SQL> ed
Wrote file afiedt.buf
  1  select empno, ename, sal
  2  from (
  3        select empno, ename, sal, dense_rank() over (partition by 1 order by sal desc) as r from emp
  4       )
  5* where r <= 3
SQL> /
     EMPNO ENAME             SAL
      7839 KING             5000
      7788 SCOTT            3000
      7902 FORD             3000
      7566 JONES            2975Using just the MAX function alone will make it harder for you and is completely unnecessary.

Similar Messages

  • How to find the greatest hiredate in emp table

    hi,
    how can I find out whose the latest recruit in the demo emp
    table.
    thanx.
    Sreekant

    Hello,
    Use following query.
    select * from emp
    where hiredate = (select max(hiredate) from emp);
    Adi

  • How to find the avg salary of first 3 employees based on each department?

    how to find the avg salary of first 3 employees based on each department?

    Hi,
    Do you mean something like this?
    with a as
    select
      deptno
      ,sal
      ,row_number() over ( partition by deptno order by sal desc) rn
    from
      emp
    select
      deptno
      ,round(avg(sal)) avg_sal_of_top_3
    from
      a
    where
      rn <= 3
    group by
      deptno
    order by
      deptno
    DEPTNO AVG_SAL_OF_TOP_3
        10             2917
        20             2992
        30             1983
    If not please give more details.
    Regards,
    Peter

  • How to find top 10  SQL statments which are consuming more cpu time.

    hi all,
    Is there any command or script to monitor the top 10 sql statments which are consuming more cpu time.
    I know by using AWR REPORT we can find it, i want the command or script to find the top cpu utilization sql statments.
    Regards
    Subhash.

    Subhash,
    A quick and dirty Google search could have get you started with the following:
    Thread: how to get top CPU consuming sql oracle 10g
    Re: how to get top CPU consuming sql oracle 10g
    Oracle SQL top sessions
    http://www.dba-oracle.com/oracle10g_tuning/t_sql_top_sessions.htm
    "How to Find top 10 expensive sql's", version 9.2.0
    http://asktom.oracle.com/pls/asktom/f?p=100:11:0::::P11_QUESTION_ID:73325450402303
    HTH,
    Thierry

  • Can you please tell me how to find top 10 wait events

    Hi
    Can you please tell me how to find top 10 wait events and what actions need to be taken when there is a wait?
    Thanks
    Regards,
    RJ.

    hi,
    suggest you to use statspack !!!!!!! for the all tuning..else use the views
    * v$session_event
    * v$session_wait
    * v$system_event
    go through this for tuning tips
    http://www.dba-oracle.com/art_dbazine_waits.htm
    Thanks
    --Raman                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                   

  • How to find top utilized query for last two months in oem

    how to find top utilized query for last two months in oracle enterprise manager?

    Can you mark the thread as Helpful  and once marked the information can be reviewed by other customer for similar queries
    Regards
    Krishnan

  • How to find top items in alv similar to MC.5

    hi abapers
    can u plz tell me how to find top items in alv grid program as we find similarly in MC.5 top material...etc as top items.
    Thanks in advance

    Nagrenda -> http://sap.ittoolbox.com/groups/technical-functional/sap-abap/how-to-display-an-alv-grid-using-abap-oo-2016165
    I suggest you read the rules of this forum.

  • How to find the level of each child table in a relational model?

    Earthlings,
    I need your help and I know that, 'yes, we can change'. Change this thread to a answered question.
    So: How to find the level of each child table in a relational model?
    I have a relacional database (9.2), all right?!
         O /* This is a child who makes N references to each of the follow N parent tables (here: three), and so on. */
        /↑\ Fks
       O"O O" <-- level 2 for first table (circle)
      /↑\ Fks
    "o"o"o" <-- level 1 for middle table (circle)
       ↑ Fk
      "º"Tips:
    - each circle represents a table;
    - red tables no have foreign key
    - the table in first line of tree, for example, has level 3, but when 3 becomes N? How much is N? This's the question.
    I started thinking about the following:
    First I have to know how to take the children:
    select distinct child.table_name child
      from all_cons_columns father
      join all_cons_columns child
    using (owner, position)
      join (select child.owner,
                   child.constraint_name fk,
                   child.table_name child,
                   child.r_constraint_name pk,
                   father.table_name father
              from all_constraints father, all_constraints child
             where child.r_owner = father.owner
               and child.r_constraint_name = father.constraint_name
               and father.constraint_type in ('P', 'U')
               and child.constraint_type = 'R'
               and child.owner = 'OWNER') aux
    using (owner)
    where child.constraint_name = aux.fk
       and child.table_name = aux.child
       and father.constraint_name = aux.pk
       and father.table_name = aux.father;Thinking...
    Let's Share!
    My thanks in advance,
    Philips
    Edited by: BluShadow on 01-Apr-2011 15:08
    formatted the code and the hierarchy for readbility

    Justin,
    Understood.
    Nocycle not work in 9.2 and, even that would work, would not be appropriate.
    With your help, I decided a much simpler way (but there is still a small problem, <font color=red>IN RED</font>):
    -- 1
    declare
      type udt_roles is table of varchar2(30) index by pls_integer;
      cRoles udt_roles;
    begin
      execute immediate 'create user philips
        identified by philips';
      select granted_role bulk collect
        into cRoles
        from user_role_privs
       where username = user;
      for i in cRoles.first .. cRoles.count loop
        execute immediate 'grant ' || cRoles(i) || ' to philips';
      end loop;
    end;
    -- 2
    create table philips.root1(root1_id number,
                               constraint root1_id_pk primary key(root1_id)
                               enable);
    grant all on philips.root1 to philips;
    create or replace trigger philips.tgr_root1
       before delete or insert or update on philips.root1
       begin
         null;
       end;
    create table philips.root2(root2_id number,
                               constraint root2_id_pk primary key(root2_id)
                               enable);
    grant all on philips.root2 to philips;
    create or replace trigger philips.tgr_root2
       before delete or insert or update on philips.root2
       begin
         null;
       end;
    create table philips.node1(node1_id number,
                               root1_id number,
                               node2_id number,
                               node4_id number,
                               constraint node1_id_pk primary key(node1_id)
                               enable,
                               constraint n1_r1_id_fk foreign key(root1_id)
                               references philips.root1(root1_id) enable,
                               constraint n1_n2_id_fk foreign key(node2_id)
                               references philips.node2(node2_id) enable,
                               constraint n1_n4_id_fk foreign key(node4_id)
                               references philips.node4(node4_id) enable);
    grant all on philips.node1 to philips;
    create or replace trigger philips.tgr_node1
       before delete or insert or update on philips.node1
       begin
         null;
       end;
    create table philips.node2(node2_id number,
                               root1_id number,
                               node3_id number,
                               constraint node2_id_pk primary key(node2_id)
                               enable,
                               constraint n2_r1_id_fk foreign key(root1_id)
                               references philips.root1(root1_id) enable,
                               constraint n2_n3_id_fk foreign key(node3_id)
                               references philips.node3(node3_id) enable);
    grant all on philips.node2 to philips;
    create or replace trigger philips.tgr_node2
       before delete or insert or update on philips.node2
       begin
         null;
       end;                          
    create table philips.node3(node3_id number,
                               root2_id number,
                               constraint node3_id_pk primary key(node3_id)
                               enable,
                               constraint n3_r2_id_fk foreign key(root2_id)
                               references philips.root2(root2_id) enable);
    grant all on philips.node3 to philips;
    create or replace trigger philips.tgr_node3
       before delete or insert or update on philips.node3
       begin
         null;
       end;                          
    create table philips.node4(node4_id number,
                               node2_id number,
                               constraint node4_id_pk primary key(node4_id)
                               enable,
                               constraint n4_n2_id_fk foreign key(node2_id)
                               references philips.node2(node2_id) enable);
    grant all on philips.node4 to philips;
    create or replace trigger philips.tgr_node4
       before delete or insert or update on philips.node4
       begin
         null;
       end;                          
    -- out of the relational model
    create table philips.node5(node5_id number,
                               constraint node5_id_pk primary key(node5_id)
                               enable);
    grant all on philips.node5 to philips;
    create or replace trigger philips.tgr_node5
       before delete or insert or update on philips.node5
       begin
         null;
       end;
    -- 3
    create table philips.dictionary(table_name varchar2(30));
    insert into philips.dictionary values ('ROOT1');
    insert into philips.dictionary values ('ROOT2');
    insert into philips.dictionary values ('NODE1');
    insert into philips.dictionary values ('NODE2');
    insert into philips.dictionary values ('NODE3');
    insert into philips.dictionary values ('NODE4');
    insert into philips.dictionary values ('NODE5');
    --4
    create or replace package body philips.pck_restore_philips as
      procedure sp_select_tables is
        aExportTablesPhilips     utl_file.file_type := null; -- file to write DDL of tables   
        aExportReferencesPhilips utl_file.file_type := null; -- file to write DDL of references
        aExportIndexesPhilips    utl_file.file_type := null; -- file to write DDL of indexes
        aExportGrantsPhilips     utl_file.file_type := null; -- file to write DDL of grants
        aExportTriggersPhilips   utl_file.file_type := null; -- file to write DDL of triggers
        sDirectory               varchar2(100) := '/app/oracle/admin/tace/utlfile'; -- directory \\bmduhom01or02 
        cTables                  udt_tables; -- collection to store table names for the relational depth
      begin
        -- omits all referential constraints:
        dbms_metadata.set_transform_param(dbms_metadata.session_transform, 'REF_CONSTRAINTS', false);
        -- omits segment attributes (physical attributes, storage attributes, tablespace, logging):
        dbms_metadata.set_transform_param(dbms_metadata.session_transform, 'SEGMENT_ATTRIBUTES', false);
        -- append a SQL terminator (; or /) to each DDL statement:
        dbms_metadata.set_transform_param(dbms_metadata.session_transform, 'SQLTERMINATOR', true);
        -- create/open files for export DDL:
        aExportTablesPhilips := utl_file.fopen(sDirectory, 'DDLTablesPhilips.pdc', 'w', 32767);
        aExportReferencesPhilips := utl_file.fopen(sDirectory, 'DDLReferencesPhilips.pdc', 'w', 32767);
        aExportIndexesPhilips := utl_file.fopen(sDirectory, 'DDLIndexesPhilips.pdc', 'w', 32767);
        aExportGrantsPhilips := utl_file.fopen(sDirectory, 'DDLGrantsPhilips.pdc', 'w', 32767);
        aExportTriggersPhilips := utl_file.fopen(sDirectory, 'DDLTriggersPhilips.pdc', 'w', 32767);
        select d.table_name bulk collect
          into cTables -- collection with the names of tables in the schema philips
          from all_tables t, philips.dictionary d
         where owner = 'PHILIPS'
           and t.table_name = d.table_name;
        -- execution
        sp_seeks_ddl(aExportTablesPhilips,
                     aExportReferencesPhilips,
                     aExportIndexesPhilips,
                     aExportGrantsPhilips,
                     aExportTriggersPhilips,
                     cTables);
        -- closes all files
        utl_file.fclose_all;
      end sp_select_tables;
      procedure sp_seeks_ddl(aExportTablesPhilips     in utl_file.file_type,
                             aExportReferencesPhilips in utl_file.file_type,
                             aExportIndexesPhilips    in utl_file.file_type,
                             aExportGrantsPhilips     in utl_file.file_type,
                             aExportTriggersPhilips   in utl_file.file_type,
                             cTables                  in out nocopy udt_tables) is
        cDDL       clob := null; -- colletion to save DDL
        plIndex    pls_integer := null;
        sTableName varchar(30) := null;
      begin
        for i in cTables.first .. cTables.count loop
          plIndex    := i;
          sTableName := cTables(plIndex);
           * Retrieves the DDL and the dependent DDL into cDDL clob       *      
          * for the selected table in the collection, and writes to file.*
          begin
            cDDL := dbms_metadata.get_ddl('TABLE', sTableName, 'PHILIPS');
            sp_writes_ddl(aExportTablesPHILIPS, cDDL);
          exception
            when dbms_metadata.object_not_found then
              null;
          end;
          begin
            cDDL := dbms_metadata.get_dependent_ddl('REF_CONSTRAINT', sTableName, 'PHILIPS');
            sp_writes_ddl(aExportReferencesPhilips, cDDL);
          exception
            when dbms_metadata.object_not_found2 then
              null;
          end;
          begin
            cDDL := dbms_metadata.get_dependent_ddl('INDEX', sTableName, 'PHILIPS');
            sp_writes_ddl(aExportIndexesPhilips, cDDL);
          exception
            when dbms_metadata.object_not_found2 then
              null;
          end;
          begin
            cDDL := dbms_metadata.get_dependent_ddl('OBJECT_GRANT', sTableName, 'PHILIPS');
            sp_writes_ddl(aExportGrantsPhilips, cDDL);
          exception
            when dbms_metadata.object_not_found2 then
              null;
          end;
          begin
            cDDL := dbms_metadata.get_dependent_ddl('TRIGGER', sTableName, 'PHILIPS');
            sp_writes_ddl(aExportTriggersPhilips, cDDL);
          exception
            when dbms_metadata.object_not_found2 then
              null;
          end;
        end loop;
      end sp_seeks_ddl;
      procedure sp_writes_ddl(aExport in utl_file.file_type,
                              cDDL    in out nocopy clob) is
        pLengthDDL  pls_integer := length(cDDL);
        plQuotient  pls_integer := null;
        plRemainder pls_integer := null;
      begin
          * Register variables to control the amount of lines needed   *
         * for each DDL and the remaining characters to the last row. *
        select trunc(pLengthDDL / 32766), mod(pLengthDDL, 32766)
          into plQuotient, plRemainder
          from dual;
          * Join DDL in the export file.                            *
         * ps. 32766 characters + 1 character for each line break. *
        -- if the size of the DDL is greater than or equal to limit the line ...
        if plQuotient >= 1 then
          -- loops for substring (lines of 32766 characters + 1 break character):
          for i in 1 .. plQuotient loop
            utl_file.put_line(aExport, substr(cDDL, 1, 32766));
            -- removes the last line, of clob, recorded in the buffer:
            cDDL := substr(cDDL, 32767, length(cDDL) - 32766);
          end loop;
        end if;
          * If any remains or the number of characters is less than the threshold (quotient = 0), *
         * no need to substring.                                                                 *
        if plRemainder > 0 then
          utl_file.put_line(aExport, cDDL);
        end if;
        -- record DDL buffered in the export file:
        utl_file.fflush(aExport);
      end sp_writes_ddl;
    begin
      -- executes main procedure:
      sp_select_tables;
    end pck_restore_philips;<font color="red">The problem is that I still have ...
    When creating the primary key index is created and this is repeated in the file indexes.
    How to avoid?</font>

  • How to get second maximum salary from employee table(sql query)

    how to get second maximum salary from employee table(sql query)

    dude there is no matter of structure .........that user already said its from employee table ...............its basic table in sql and there is no need to specify the table structure
    .........i think u got my point I think you are the one who didn't understand Sarma's point.
    Give a man a fish and you feed him once. Teach a man how to fish and you feed him a life long.
    >
    and the query is
    select max(sal) from emp where sal<(select max(sal)
    from emp);
    this will give the 2nd max salary from the emp tableBtw: You solution is bad, because it needs to scan and sort the table emp twice. And a better solution has been given already.
    Message was edited by:
    Sven W. - reordered statements

  • To find the lowest salary from a table

    hello guys,
    i need to find the lowest salary from a table. whats the easy way to do this.

    How do you handle ties?
    If you want multiple rows returned when there are multiple people with the same salary
    SELECT *
      FROM(
        SELECT a.*,
               rank() over (order by salary desc ) rnk
          FROM your_table a)
    WHERE rnk = 1If you want to handle ties differently, use a different analytic function. Probably dense_rank or row_number.
    Justin

  • How to find total recs in a local table for a particular condition

    Hi,
    How to find total recs in a local table for a particular condition?
    Thanks,
    CD

    Well, you may want to try this as well, and compare to the LOOP way.  Not sure what kind of overhead you may get doing this way. Here ITAB is our main internal table, and ITAB_TMP is a copy of it.  Again I think there may be some overhead in doing the copy.  Next, delete out all records which are the reverse of your condition.  Then whatever is left is the rows that you want to count.  Then simply do a LINES operator on the internal table, passing the number of lines to LV_COUNT.
    data: itab type table of ttab.
    data: itab_tmp type table of ttab.
    itab_tmp[] = itab[].
    delete table itab_tmp where fld1 <> 'A'.
    lv_count = lines( itab_tmp ).
    Regards,
    Rich Heilman

  • How to find out the selected column in Table Control

    Hi all,
          How to find out the selected column in Table Control?
    Thanks & Regards,
    YJR

    Hi,
    Let your table control name in Screen painter be TC1.
    READ TABLE TC1-COLS INTO WA_COLS (some wok area)
                 WITH KEY SELECTED = 'X'.
            IF SY-SUBRC = 0.
              CLEAR: W_DUMMY, W_COL_NAME.
              SPLIT WA_COLS-SCREEN-NAME AT '-' INTO W_DUMMY
                                                   W_COL_NAME.
            endif.
    W_COL_NAME gives you the column name.
    Hope it helps.
    cheers
    sharmistha

  • How to find out the link for the tables for sub contract PO

    Dear All,
    How to find out the link for the tables for sub contract PO, I am looking for the fields Material, Unit, Batch no., quantity.
    Please provide the solution for the Table related to ABAP.
    Thanks,
    V.V.

    Hi,
       Make use of the table EKKO, EKPO...
    For subcontract please select the option in EKPO-PSTYP.
    Regards
    GK.

  • How to find the structural difference between two tables

    Hi all,
    How to find the structural difference between two tables .
    Oracle Database 11g Enterprise Edition Release 11.1.0.7.0 - Production
    PL/SQL Release 11.1.0.7.0 - Production
    CORE 11.1.0.7.0 Production
    TNS for 32-bit Windows: Version 11.1.0.7.0 - Production
    NLSRTL Version 11.1.0.7.0 - Production
    Thanks,
    P Prakash

    you could try something similar to this, for each table pair that you want to compare:
    SELECT 'TABLE_A has these columns that are not in TABLE_B', DIFF.*
      FROM (
            SELECT  COLUMN_NAME, DATA_TYPE, DATA_LENGTH
              FROM all_tab_columns
             WHERE table_name = 'TABLE_A'
             MINUS
            SELECT COLUMN_NAME, DATA_TYPE, DATA_LENGTH
              FROM all_tab_columns
             WHERE table_name = 'TABLE_B'
          ) DIFF
    UNION
    SELECT 'TABLE_B has these columns that are not in TABLE_A', DIFF.*
      FROM (
            SELECT COLUMN_NAME, DATA_TYPE, DATA_LENGTH
              FROM all_tab_columns
             WHERE table_name = 'TABLE_B'
             MINUS
            SELECT COLUMN_NAME, DATA_TYPE, DATA_LENGTH
              FROM all_tab_columns
             WHERE table_name = 'TABLE_A'
          ) DIFF;that's assuming, column_name, data_type and data_length are all you want to compare on.

  • How to find last inserted record in the table.

    Version: Oracle 10g
    I have a table called "Manufacture" and 3 columns as mfno,itemname,quantity.
    How to find last inserted record in the table "Manufacture".
    As i come to know that Rowid is not result perfect results. Please provide your inputs.

    user13416294 wrote:
    Version: Oracle 10gThat is not a version. That is a product name. A version is 10.1.0.2 or 10.2.0.4, etc.
    I have a table called "Manufacture" and 3 columns as mfno,itemname,quantity.
    How to find last inserted record in the table "Manufacture".Not possible as your data model does not cater for it. That simple.
    If there is a need to determine some order or associate some time to an entity, then that should be part of the data model - and a relationship, or one or more attributes are needed to represent that information. Thus your data model in this case is unable to meet your requirements.
    If the requirements are valid, fix the data model. In other words - your question has nothing to do with Oracle and nothing to do with rowid, rowscn or other pseudo columns in Oracle. It is a pure data modeling issue. Nothing more.

Maybe you are looking for