SQL help (line column... a little)

Similar Questions

• SQL help - lines, view by grouping...

  Greetings!
  
  Is this possible in Oracle 10 g using the SQL statement.
  
  EMP table - EMPLID, DEPTID fields.
  Data: E01, D01
  E02, D01
  E03, D01
  E11, D11
  E12, D11
  
  I'm looking for output as follows using a SQL statement...
  
  DeptID, EmplID
  D01 E01
  E02
  E03
  D11, E11
  E12
  
  
  Thanks in advance.

  Try this:

  WITH emp AS
  (SELECT 'E01' emplid, 'D01' deptid FROM dual UNION ALL
   SELECT 'E02' emplid, 'D01' deptid FROM dual UNION ALL
   SELECT 'E03' emplid, 'D01' deptid FROM dual UNION ALL
   SELECT 'E11' emplid, 'D11' deptid FROM dual UNION ALL
   SELECT 'E12' emplid, 'D11' deptid FROM dual)
  SELECT   CASE ROW_NUMBER () OVER (PARTITION BY deptid ORDER BY deptid)
           WHEN 1 THEN deptid||','||emplid
           ELSE emplid
           END dept_emp
  FROM     emp
  ORDER BY deptid,
           emplid
  

• Database trigger - PL/SQL: ORA-00984: column not allowed here

  I am trying to create a trigger that will update a table of audit used when a row is changed. Using a sequence number to assign an identifier unique to each line as it is created. Need to capture the user ID, date modified and action (update), the image of the front line.
  

  CREATE SEQUENCE emp_audit_seq START WITH 10;                

  Create table emp (
     empno       NUMBER(4)      Primary Key,
     ename       VARCHAR2(10),
     job            VARCHAR2(9),
     mgr           NUMBER(4),
     hiredate     DATE,
     sal             NUMBER(7,2),
     comm        NUMBER(7,2),
     deptno       NUMBER(2));

  CREATE TABLE emp_audit   (
       audit_uid          NUMBER(15)      Primary Key,
       change_date          DATE,
       change_user          VARCHAR2(30),
       action                  CHAR(1),
       empno                  NUMBER(4),
       ename                  VARCHAR2(10),          
       job               VARCHAR2(9),
       mgr               NUMBER(4),
       hiredate          DATE,
       sal               NUMBER(7,2),
       comm                  NUMBER(7,2),
       deptno                  NUMBER(2));

  CREATE OR REPLACE TRIGGER trig_emp_audit
    BEFORE UPDATE ON emp
    FOR EACH ROW
  BEGIN
    INSERT INTO emp_audit
      VALUES(emp_audit_seq.nextval, change_date, change_user, action, :old.empno, :old.ename, :old.job, :old.mgr, :old.hiredate, :old.sal, :old.comm, deptno);
  END;
  /
  
  Warning: Trigger created with compilation errors.
  
  SQL> show errors
  Errors for TRIGGER TRIG_EMP_AUDIT:
  
  LINE/COL ERROR
  -------- -----------------------------------------------
  2/3      PL/SQL: SQL Statement ignored
  3/149    PL/SQL: ORA-00984: column not allowed here

  Can someone help to help me find what I'm doing wrong with the trigger?
  
  Published by: LostNoob on August 25, 2012 14:24

  First of all, when you write an INSERT statement, it is always good for the columns that you insert in the list. Which makes the code easier to follow - you do not have separately pull toward the top of the table definition to know what order of columns is inserted. And it makes the code easier to manage because the declaration become invalid if you add a new column to the table in the future.

  Second, CHANGE_DATE, CHANGE_USER and ACTION are not (probably) functions and are not local variables so it is not supposed to use them in an INSERT statement. You need to write code or to take advantage of the existing functions to fill in these columns. I suppose, for example, that you want to use SYSDATE to fill the CHANGE_DATE and the USER to fill the column CHANGE_USER. My guess is that ACTION must always be a 'U' for UPDATE.

  Thirdly, it seems that you left the: old man on the DEPTNO column.

  Put them all together, you would have something like

  CREATE OR REPLACE TRIGGER trig_emp_audit
    BEFORE UPDATE ON emp
    FOR EACH ROW
  BEGIN
    INSERT INTO emp_audit(
        audit_uid,
        change_date,
        change_user,
        action,
        enpno,
        ename,
        job,
        mgr,
        hiredate,
        sal,
        comm,
        deptno )
      VALUES(
        emp_audit_seq.nextval,
        sysdate,
        user,
        'U',
       :old.empno,
       :old.ename,
       :old.job,
       :old.mgr,
       :old.hiredate,
       :old.sal,
       :old.comm,
       :old.deptno);
  END;
  / 
  

  Justin

• PL/SQL: ORA-00984: column not allowed here

  Sorry, it's probably easy and I forgot something simple, but it's driving me crazy :-)
  

  VARIABLE g_fk_deduction      VARCHAR2(30)
  VARIABLE g_fk_empno          NUMBER
  VARIABLE g_before_or_after_flag     CHAR(1)
  VARIABLE g_deduction_amount     NUMBER
  
  BEGIN
    :g_fk_deduction           := '401K';
    :g_fk_empno               := 7369;
    :g_before_or_after_flag     := 'B';
    :g_deduction_amount          := 150.00;
  END;
  /
  BEGIN
    INSERT INTO emp_deductions      (fk_deduction, fk_empno, before_or_after_flag, deduction_amount)
     VALUES               (g_fk_deduction, g_fk_empno, g_before_or_after_flag, g_deduction_amount);
    COMMIT;
  END;
  /

  Error: PL/SQL: ORA-00984: column not allowed here on g_deduction_amount
  in the value clause.
  Any help would be appreciated.
  
  Table is below:

  CREATE TABLE emp_deductions     (
        fk_deduction            VARCHAR2(30),
        fk_empno                  NUMBER(4),
        before_or_after_flag         CHAR(1),
        deduction_amount            NUMBER(6,2));

  Published by: LostNoob on August 23, 2012 19:06

  rp0428 wrote:
  >
  : g_fk_deduction: = "401k".
  : g_fk_empno: = 7369;
  : g_before_or_after_flag: = 'B ';.
  : g_deduction_amount: = 150.00;
  >
  Why did you put a colon here? Get rid of them.

  They are necessary, since they are declared SQL * more variable.

  The problem for the OP, is that in the clause values in the insert the colon are missing.

  Published by: Mark Williams on August 23, 2012 22:31

  Here is your example requested:

  SQL> create table test (c number);
  
  Table created.
  
  SQL> variable v_c number
  SQL> begin
    2    :v_c := 46;
    3  end;
    4  /
  
  PL/SQL procedure successfully completed.
  
  SQL> begin
    2    insert into test values (:v_c);
    3  end;
    4  /
  
  PL/SQL procedure successfully completed.
  
  SQL> commit;
  
  Commit complete.
  
  SQL> select * from test;
  
           C
  ----------
          46
  
  1 row selected.
  
  SQL>
  

• Add the elements of a single line/column of a matrix

  Hi all

  What is the best way to add all the elements of a single row or column of a given matrix the matrix and the number of this line/column on LabVIEW? The output should be a number for each row or column.

  ARIJ,

  You always said not if you use a table or a matrix.

  The transposition has to be out of the loop.

  This will give you the sum of all rows and all columns.

  These are functions and LabVIEW base constructions. If you do not understand the concepts, please visit the online tutorials.

  Lynn

  

• Added line/column in the analysis document Web

  Dear,

  I am a beginner in Web Analytics. Is it possible to add a line/column of text/formula in a document?

  I use version 11.1.2.3

  Kind regards

  Ahmad.

  Hello

  To insert a row or a column in a grid of free-form, with the right button of the freeform grid, and then select Insert a line or insert a column

  HTH

  Kind regards

  Zohra.

• Try to install CS4 Design Premium - disabled since the old computer, new computer does not accept my serial number.  Just wasted 2 hours on Adobe help line.

  Try to install CS4 Design Premium - disabled since the old computer, new computer does not accept my serial number.  Just wasted 2 hours on Adobe help line.

  I sent you a private message that will help you (I hope).

  because scammers can continue to check this thread, I don't want to enter here information that could help them and this is the reason why I deleted your last post and sent you a private message.

• I just downloaded the new DC Acrobat, and I'm having a lot of problems with it. It's very, very slow. He has opening, saving, closing, do anything. Is there something that can be done to help him run a little more, well, normally?

  I just downloaded the new DC Acrobat, and I'm having a lot of problems with it. It's very, very slow. He has opening, saving, closing, do anything. Is there something that can be done to help him run a little more, well, normally?

  I fixed it! I used the string analyze function waiting in the Task Manager while opening a PDF file and have to wait 12 seconds... and it showed that the wait chain has been blocked by the Speech SDK. So I went to my Plugins of PDF from Adobe folder, which in my case is C:\Program Files (x 86) \Adobe\Acrobat DC\Acrobat\plug_ins and delete the plugin ReadOutLoud.api. (I had to close all instances of Acrobat DC first, and I had to also click the UAC Popup Windows application approval raised to remove the plugin) Now... instant Adobe Acrobat DC charge!

• SQL monitor help with column of chronology

  

  In the report attached to sql, in the timeline column means the whole last 3 began simultaneously and hash join is the first exploited who started?

  my understanding is 'index range scan' should start first and give the data to "Single Partition range". Only once this step completed, he should 'TABLE ACCESS Full'. Isn't this correct?

  Thanks for your time!

  Your description is correct - and I think that you just have to allow some defects in the rounded, graphic display of errors and inconsistencies in what is "the order of execution.

  If it is still possible, you can watch dbms_sqltune.report_sql_monitor output text () as numbers COULD clarify the sequence - on the other hand, the granularity of the timer is not very fine.

  You may find that the chronology of hash join begins before the access times two table - which highlights the problem of "means to launch an operation."  The subroutine "do a hash join" must begin before the two tablescans, on the other hand the mechanism 'return rowsource to parent' cannot start until after the second table access began to return the hash join lines.

  Another sign of the available seeds, of course, is that the select statement does not seem to have started until a few seconds after the hash join has started!

  Concerning

  Jonathan Lewis

• SQL Help: convert values in two rows in a line

  Hello
  
  I am using Oracle 11.2, I want to get the values in different rows in the Tb_a table and insert them into the TB_b table as suite:
  
  create a Tb_a (number (5) id, init_dt date, upd_dt date, addr varchar2 (32));
  create a Tb_b (number (5) id, date upd_dt, addr varchar2 (32), dated hist_dt);
  
  insert into Tb_a values (101, to_date('ddmmyyyy','21/08/1990'), to_date('ddmmyyyy','22/08/1990'), "address1");
  insert into Tb_a values (101, to_date ('ddmmyyyy', 21 August 1991 '), to_date ('ddmmyyyy', 22 August 1991 '), "address2");
  insert into Tb_a values (101, to_date('ddmmyyyy','21/08/1992'), to_date('ddmmyyyy','22/08/1992'), "address3");
  insert into Tb_a values (102, to_date('ddmmyyyy','24/08/1990'), to_date('ddmmyyyy','25/08/1990'), 'address_address');
  
  I would like to have data in TB_b as follows.
  101, 22 August 1990 ","address1", August 21, 1991"
  101, August 22, 1991 ","address2", August 21, 1992 '"
  101, August 22, 1992 ', 'address3', null
  102, 25 August 1990 ', 'address_address', null
  
  Suggest on the SQL to perform the conversion of Tb_a to Tb_b?
  
  Thank you

  939569 wrote:
  I wish that each line of TB_b has the same value to the ID of the column, upd_dt and addr as Tb_a and the next value in the crt_dt of the Tb_a column.

  Or table has a column called crt_dt. Do you mean init_dt?

  This rule applies to the same ID only. How can I use LEAD on different ID, not the whole picture?

  Inlcude

  PARTITION BY id
  

  in the analytical clause, like this:

  INSERT INTO  tb_b ( id
                          , upd_dt
              , addr
              , hist_dt
              )
  SELECT  id
  ,     init_dt
  ,     addr
  ,     LEAD (init_dt) OVER ( PARTITION BY  id
                                   ORDER BY         init_dt
                     )
  FROM    tb_a;
  

• SQL Help: Derive from 'News' columns based on condition

  Hi all
  
  Can someone help me with this please? I am trying to achieve in a SQL (no), if this can be done in two different stages, even that is fine too...
  
  Thanks in advance,
  Chavigny
  

  SELECT * FROM
  (
  SELECT 2765 BTCH_ID,'payroll' BTCH_NM,'payroll~EUR~EUR~DE'NEW_BTCH_NM,'1234'TMPLT_ID FROM DUAL
  UNION
  SELECT 2765,'payroll','payroll~USD~USD~US','1235' FROM DUAL
  UNION
  SELECT 2765,'payroll','payroll~USD~USD~US','1236' FROM DUAL
  UNION
  SELECT 2765,'payroll','payroll~GBP~GBP~GB','1237' FROM DUAL
  UNION
  SELECT 2766,'SALARY','SALARY~USD~USD~US','1238' FROM DUAL
  UNION
  SELECT 2766,'SALARY','SALARY~USD~USD~US','1239' FROM DUAL
  ) T_BATCH;

  For a "batch - id", there are several Transactions that are denoted by (cr_crncy_cd, dr_crncy_cd, recvr_cntry_cd), and the "NEW_BTCH_NM" column is basically a concatenation of 'BTCH_NM' + above 3 columns (with a ~)
  
  I'm trying to derive two new columns "SEQ" & "Output_Batch_name".
  
  The SEQ column:
  

  For every batch_id, the first row will be assigned by sequence.nextval.
      For all transactions within BTCH_ID, 
                     all the rows with same NEW_BTCH_NM, share the same sequence number
                      if different, then SEQ.nextval for all distinct rows

  Output_BTCH_NM:
  

     for a given batch_id, 
           if the NEW_BTCH_NM is the same for all transactions ( batch_id 2766 in the below example) ,  
                           then Output_BTCH_NM = BTCH_NM should be marked for All transactions which belong to this Batch_id,
          Else 
              Output_BTCH_NM = New_BTCH_NM
       END if

  EXPECTED RESULTS
  

  *BTCH_ID     BTCH_NM             NEW_BTCH_NM          TMPLT_ID     SEQ         Output_BTCH_NM*
  2765       payroll                payroll~EUR~EUR~DE     1234         1          payroll~EUR~EUR~DE
  2765       payroll                payroll~GBP~GBP~GB     1237         2          payroll~GBP~GBP~GB
  2765       payroll                payroll~USD~USD~US     1235         3          payroll~USD~USD~US
  2765       payroll                payroll~USD~USD~US     1236         3          payroll~USD~USD~US
  
  2766       SALARY                SALARY~USD~USD~US     1238         4          SALARY
  2766       SALARY                SALARY~USD~USD~US     1239         4          SALARY

  Published by: user10711957 on 25 Sep, 2010 16:37

  An extension of the solution of the (slightly modified) jiq

  function my_sequence(whichval in varchar2) return number is
  begin
    if whichval = 'currval' then
      return your_sequence.currval;
    else
      return your_sequence.nextval;
    end if;
  end my_sequence;
  
  WITH t AS
  (
  SELECT 2765 BTCH_ID,'payroll' BTCH_NM,'payroll~EUR~EUR~DE'NEW_BTCH_NM,'1234'TMPLT_ID FROM DUAL
  UNION
  SELECT 2765,'payroll','payroll~USD~USD~US','1235' FROM DUAL
  UNION
  SELECT 2765,'payroll','payroll~USD~USD~US','1236' FROM DUAL
  UNION
  SELECT 2765,'payroll','payroll~GBP~GBP~GB','1237' FROM DUAL
  UNION
  SELECT 2766,'SALARY','SALARY~USD~USD~US','1238' FROM DUAL
  UNION
  SELECT 2766,'SALARY','SALARY~USD~USD~US','1239' FROM DUAL
  )
  select btch_id,btch_nm,new_btch_nm,tmplt_id,
         case when lag(seq,1) over (order by tmplt_id) = seq
              then my_sequence('currval')
              else my_sequence('nextval')
         end seq,
         out_btch_nm
  select btch_id,btch_nm,new_btch_nm,tmplt_id,
         dense_rank() over(partition by btch_id order by new_btch_nm) seq
         case when count(distinct new_btch_nm) over (partition by btch_id) = 1
              then btch_nm
              else new_btch_nm
         end out_btch_nm
    from t
  

  Concerning

  Etbin

  Posted before you see the following

  the sequence # must be continuous, must avoid any shortfall or gaps with the numbering.

  the foregoing must be the only user of its own sequence somehow and maybe...
  Search for "sequence free gap"... http://asktom.oracle.com/pls/asktom/f?p=100:11:0:P11_QUESTION_ID:4343369880986
  Edited by: Etbin on 26.9.2010 08:21

• SQL puzzler: Line rankings according to the columns

  We have a table that looks like the following:

  COL_1_RANK	COL_1_VALUE	COL_2_RANK	COL_2_VALUE
  1	AAA	3	HHH
  2	BBB	2	GGG
  3	CCC	5	JJJ
  4	DDD	1	FFF
  5	EEE	4	III

  We want to extract a *single* rank for each case where COL_1_RANK and COL_2_RANK are equivalent. Then we want the values corresponding to this mutual rank to be contained in the same line. Thus, the desired output is as follows:

  MUTUAL_RANK	COL_1_VALUE	COL_2_VALUE
  1	AAA	FFF
  2	BBB	GGG
  3	CCC	HHH
  4	DDD	III
  5	EEE	JJJ

  The other condition is that we want to do this in a *table single pass*. So we are aware of the solutions using two different CLAUSES and by joining in the ranks. It's not what we're looking for here.

  We are suspecting that some analytic function can be used here, but can't seem to find, he. Basically, for each line, take the value of COL_1_RANK and, then, on the set of all values of COL_2_RANK , locate the line where COL_2_RANK = s COL_1_RANK. Pull the COL_2_VALUE of this line.

  No matter what SQL Smarties out there who have a solution for this?

  Thank you

  -Joe

  SELECT col_1_rank,

  col_1_value,

  REGEXP_SUBSTR (listagg (col_2_value, ",") THE Group (ORDER BY col_2_rank) (),

  '[^,]+',

  1,

  col_1_rank)

  col_2_value

  T

  order by 1;

• SQL help, how to group in column A and then get a higher frequency in column B?

  Assuming that the following table, operation represents each type of operation of transaction
  Employee table operation another column c
  ' < ' id1
  ' < ' id1
  ' < ' id2
  ' *' id3
  id1 ' / '.
  
  question, for each operation, which employee do the most? which employee do the 2nd most?
  I thought I need to group by operation first, then get the frequency of the employee and get top 1, or at the top of page 2 of higher frequency.
  But I'm struct and don't know how to write sql code.

  998408 wrote:
  Assuming that the following table, operation represents each type of operation of transaction
  Employee table operation another column c
  '<'>
  '<'>
  '<'>
  ' *' id3
  id1 ' / '.

  question, for each operation, which employee do the most? which employee do the 2nd most?
  I thought I need to group by operation first, then get the frequency of the employee and get top 1, or at the top of page 2 of higher frequency.
  But I'm struct and don't know how to write sql code.

  Welcome to OTN. OPS! Posted in the wrong forum. Just after in {forum: id = 75}. Before posting it close this thread as answered marking.

  If someone answer is useful or appropriate, please mark accordingly. *

• SQL help - Need help to rotate the columns to rows

  I have a HughesNet to divide the columns into multiple lines. For example:
  
  EMP_DEPT
  
  ROWID empid1 ename1 empid2 ename2 empid2 ename2 empid4 ename4 dept4 dep3 dep2 dept1
  100001 1 'SCOTT' 10 2 'DAVE' 20 3 10 4 20 SMITH "MILLER"
  100002 1 'SCOTT' 10 2 'DAVE' 20 3 'MILLER' 20
  
  Note: EMP_DEPT do not always have information about the 4 employees settled for example in info only 3 employees rank 2 are there
  
  I need to convert and insert it into the EMPLOYEE table as follows:
  
  EMPLOYEE
  
  EmpID ename dept
  1 SCOTT 10
  2 20 DAVE
  3 MILLER 10
  4 SMITH 20
  
  1 SCOTT 10
  2 20 DAVE
  3 MILLER 20
  
  Thank you
  KeV

  Hey Kevin,

  Here's one way:

  WITH t AS (
    SELECT level i FROM dual CONNECT BY level <= 4
  )
  SELECT enty_type, enty_name, enty_id
  FROM (
    SELECT case when mod(t.i,2) = 0 then 'DEPARTMENT'
                else 'EMPLOYEE'
           end as enty_type
         , case t.i
             when 1 then emp_name1
             when 2 then dept_name1
             when 3 then emp_name2
             when 4 then dept_name2
           end as enty_name
         , case t.i
             when 1 then emp_id1
             when 2 then dept_id1
             when 3 then emp_id2
             when 4 then dept_id2
           end as enty_id
    FROM emp
         CROSS JOIN t
  )
  WHERE enty_id IS NOT NULL
  ;
  

  Another using the MODEL clause:

  SELECT *
  FROM (
    SELECT enty_id, enty_name, enty_type
    FROM emp
    MODEL
    RETURN UPDATED ROWS
    PARTITION BY (pk)
    DIMENSION BY (0 i)
    MEASURES(
       emp_id1, emp_name1
     , emp_id2, emp_name2
     , dept_id1, dept_name1
     , dept_id2, dept_name2
     , cast(null as number(10)) enty_id
     , cast(null as varchar2(200)) enty_name
     , cast(null as varchar2(30)) enty_type
    )
    RULES (
       enty_type[1] = 'EMPLOYEE' , enty_id[1] = emp_id1[0], enty_name[1] = emp_name1[0]
     , enty_type[2] = 'EMPLOYEE' , enty_id[2] = emp_id2[0], enty_name[2] = emp_name2[0]
     , enty_type[3] = 'DEPARTMENT' , enty_id[3] = dept_id1[0], enty_name[3] = dept_name1[0]
     , enty_type[4] = 'DEPARTMENT' , enty_id[4] = dept_id2[0], enty_name[4] = dept_name2[0]
    )
  )
  WHERE enty_id IS NOT NULL
  ;
  

  Published by: odie_63 on 8 Dec. 2010 21:00

• With the help of functions/sql in the columns of the report

  Hello
  
  New to CRM OD then faced with reports...
  
  So the problem is:
  
  Have a record of account and an account of 'Parent' with her (Type: account)
  
  Now, the report should show the name of the account, the Parent Account(if any), the ID of the Parent account.
  
  Don't know how to implement the 3rd colonel IE parent account ID. It always shows the ID of the main account itself.
  
  Any suggestions will be greatly useful.
  
  Thank you!!!

  Perhaps this is an option to add a (hidden) field on the account that has the Id of the Parent account. You must fill in this field with a default value and a workflow, in the case where the parent gets later added or changed. This case, the parent Id account will be storen on registration account itself allowing you to report on this subject. It's a little workaround, but I think it works. Let me know if you need help on getting the right syntax.

  Best regards, Tim