Complex SQL help

Similar Questions

• Help needed on complex sql resolution

  Hi all
  
  I have a table such as indicated and wanted to insert these records in both tables and get the number of each type...
  

  CREATE TABLE RETURN_DEV
  (
    RCODE       VARCHAR2(4 BYTE),
    RNUMBER     VARCHAR2(16 BYTE),
    RDESC       VARCHAR2(40 BYTE),
    RDATE_TIME  VARCHAR2(16 BYTE),
    RQUANTITY   VARCHAR2(20 BYTE)
  )
  
  
  Insert into RETURN_DEV    (RCODE, RNUMBER, RDESC, RDATE_TIME, RQUANTITY)
   Values    ('0987', '919341035404586', 'POS:00003:RETR', '20100112112324', '1');
  Insert into RETURN_DEV    (RCODE, RNUMBER, RDESC, RDATE_TIME, RQUANTITY)
   Values    ('0910', '929341035404586', 'POS:00003:REMR', '20100102112324', '1');
  Insert into RETURN_DEV    (RCODE, RNUMBER, RDESC, RDATE_TIME, RQUANTITY)
   Values    ('0914', '939341035404586', 'POS:00003:REMR', '20100111112324', '1');
  Insert into RETURN_DEV    (RCODE, RNUMBER, RDESC, RDATE_TIME, RQUANTITY)
   Values    ('0997', '357991027440360', 'POS:00003: WEXC', '20100102112324', '1');
  Insert into RETURN_DEV    (RCODE, RNUMBER, RDESC, RDATE_TIME, RQUANTITY)
   Values    ('0997', '347991027440360', 'POS:00003: WEXC', '20100102112324', '1');
  Insert into RETURN_DEV    (RCODE, RNUMBER, RDESC, RDATE_TIME, RQUANTITY)
   Values    ('0987', '949341035404586', 'POS:00003:SAXP', '20100512112324', '1');
  Insert into RETURN_DEV    (RCODE, RNUMBER, RDESC, RDATE_TIME, RQUANTITY)
   Values    ('0956', '949341035404586', 'POS:00003:RETR', '20100212112324', '1');
  Insert into RETURN_DEV    (RCODE, RNUMBER, RDESC, RDATE_TIME, RQUANTITY)
   Values    ('0968', '949341035404586', 'POS:00003:SAXP', '20100312112324', '1');
  Insert into RETURN_DEV    (RCODE, RNUMBER, RDESC, RDATE_TIME, RQUANTITY)
   Values    ('0967', '997991027440360', 'POS:00003: USXP', '20060102112324', '-1');
  Insert into RETURN_DEV    (RCODE, RNUMBER, RDESC, RDATE_TIME, RQUANTITY)
   Values    ('0967', '997991027440360', 'POS:00003: USXP', '20070102112324', '1');
  Insert into RETURN_DEV    (RCODE, RNUMBER, RDESC, RDATE_TIME, RQUANTITY)
   Values    ('0967', '997991027440360', 'POS:00003: USXP', '20080102112324', '-1');
  COMMIT;
  
  Now my Algorithm go like this 
  
  step 1: Load 'RET' type into Retail_Table and Count 'RET' type
  If the RCODE  is not in [0910, 0914, 0997] and the RDESC begins with "POS:00003" and RQUANTITY =+1 load these as 'RET' type
  
  Step2: Load   'REM' type type into Retail_Table and Count 'REM' type
  If the RCODE  is in [0910, 0914] and RQUANTITY = +1, load these as 'REM' type
  
  Step3: Load  'WEX' type into Retail_Table and Count 'WEX' type
  
  If the RCODE  = 0997 and RQUANTITY = +1, load these as 'WEX' type
  
  For me main problem starts here....
  
  Step4: Load'RET' INTO  type  and 'RETXP' into retail_exceptions, and update the count of 'RET' Type and count 'RETXP' type
  
  If the RCODE  is not in [0910, 0914, 0997] and the RDESC  begins with "POS:00003" 
  and  if there is more than one record fitting that criteria in the file for the RNUMBER and the newest record chronologically (RDATE_TIME) has RQUANTITY = +1, 
  then load the latest record as Retail Returns and load all other records into retail_exceptions table.
  
  Step5: Load Retail Returns Exceptions as 'RETXP', update the count of Retail Returns  and add the count of exceptions
  
  If the RCODE  is not in [0910, 0914, 0997] and the RDESC  begins with "POS:00003" 
  and if there is more than one record fitting that criteria  for this RNUMBER and the newest record chronologically (RDATE_TIME)  has RQUANTITY = -1, 
  then load all records into to the Retail exceptions table

  I asssume we can get the account once the algorithm sucssefully inserts the date in the tables.
  

  Both the table has the same structure as shown..
  
  CREATE TABLE Retail_Table
  (
    RCODE       VARCHAR2(4 BYTE),
    RNUMBER     VARCHAR2(16 BYTE),
    RDESC       VARCHAR2(40 BYTE),
    RDATE_TIME  VARCHAR2(16 BYTE),
    RTYPE   VARCHAR2(20 BYTE)
  )
  
  
  CREATE TABLE Retail_Table_Exceptions
  (
    RCODE       VARCHAR2(4 BYTE),
    RNUMBER     VARCHAR2(16 BYTE),
    RDESC       VARCHAR2(40 BYTE),
    RDATE_TIME  VARCHAR2(16 BYTE),
    RTYPE   VARCHAR2(20 BYTE)
  )

  Thanks in advance...
  
  Published by: Rede on October 19, 2010 16:20

  Hello

  Sorry, I don't understand still.

  Rede says:
  Thanks Frank.I thought to explain clearly, but it became huge, then I'll be more clear and precise...

  I expect the output as shown below

  Output Sample data of Retail Table....
  
  RDESC            RQUANTITY     RNUMBER     TRANSACTION_TYPE
  
  POS:00003: USXP     -1     997991027440360     RET
  POS:00003: WEXC     1     347991027440360     WEX
  POS:00003: WEXC     1     357991027440360     WEX
  POS:00003:REMR     1     929341035404586     REM
  POS:00003:REMR     1     939341035404586     REM
  
  out put sample data of Retail_Exception Table
  
  RDESC            RQUANTITY     RNUMBER     TRANSACTION_TYPE
  POS:00003: USXP     -1     997991027440360     RETXP
  POS:00003: USXP     1     997991027440360     RETXP
  POS:00003:RETR     1     949341035404586     RETXP
  POS:00003:SAXP     1     949341035404586     RETXP
  

  CREATE orders TABLE for the retail and retail_exception do not have a rquantity column. If they have this column?
  Are the other columns (rcode and rdate_time) of retail and retail_exception supposed to be NULL at this point?

  What does each line in these tables?
  Why there are 9 rows in both tables combined, instead of 11?

  What i want is individaul counts of all trasaction types.If we can get the counts at time of DML querry it will be good.
  But i want to keep them in a local variable thats the reason i want to do it in a seperate sql querry.
  
  as shown below  
  
  Transaction_Type    Count
  RETXP                    4
  WEX                      2
  RET                       2
  

  Please explain how you get these results.
  Why is the number 2 when transaction_type = "RET"? Why not 1?
  Why don't you want a line to transaction_type = "REM"?

  These are all numeric values only. since i am getting these values from a external table i kept them as Varchar's.
  But i will change them accordingly.
  

  What i am trying to do is i grouping the items on the different conditions given below
  
  1. When RCODE In ('0997')            And Rquantity = '1'   Then 'WEX'
  
  2. When RCODE In ('0910', '0914') And Rquantity = '1'    Then 'REM'
  
  3. When RCODE Not  In ('0997', '0910', '0914')   And RDesc Like 'POS:00003%'
  And Count ( * ) Over (Partition By RNUMBER) = 1 Then  'RET'
  
  its easy till above...but if the third condition has more than 1 record and if the top most record has 
  
  4.    Rquantity = '-1' then categorise all the records  'RETXP' of that type   ( that satisfies the third condition)   else if top most record
  5.    Rquantity = '1' then first record should be categorised as 'RET' and the remaining of that type as 'RETXP'.
  
  top most is the record that is sorted by date for a given  RNumber.
  

  What is the "first record"? Is it the same as the 'highest '?

  Please give specific examples of how you get the desired results of the sample data that you posted.

  For example, how do you get the output line:

  RDESC               RQUANTITY  RNUMBER           TRANSACTION_TYPE
  ---------------   ---------  ---------------  ----------------
  POS:00003: USXP       -1          997991027440360  RET
  

  This line of output appears to be the 'top most' line for rnumber = '997991027440360':

  RCODE RNUMBER          RDESC              RDATE_TIME       RQUANTITY
  ----- ---------------- ------------------ ---------------- ---------
  0967  997991027440360  POS:00003: USXP    20060102112324   -1
  0967  997991027440360  POS:00003: USXP    20070102112324   1
  0967  997991027440360  POS:00003: USXP    20080102112324   -1
  

  Rule 4 does not apply to this rnumber? Isn't the line with rdate_time = '20060102112324' the line ' top more? It is not rquantity = '-1'?

  You can do the simple CASE expression (and probably faster) some of the functions of calculation only once, in a subquery, like this:

  WITH     got_analytics     AS
  (
       SELECT     rd.*
       ,     COUNT (*)           OVER ( PARTITION BY  rnumber)     AS rnumber_cnt
       ,     FIRST_VALUE (rquantity) OVER ( PARTITION BY  rnumber
                                  ORDER BY          rdate_time
                                )                    AS top_rquantity
       ,     ROW_NUMBER ()          OVER ( PARTITION BY  rnumber
                                  ORDER BY          rdate_time
                                )                    AS seq_number
       FROM     return_dev     rd
  )
  SELECT     rdesc
  ,      rquantity
  ,      rnumber
  ,     CASE
              WHEN  rcode      IN ('0997')
              AND   rquantity      = '1'               THEN     'WEX'
              WHEN  rcode           IN ('0910', '0914')
              AND   rquantity      = '1'                THEN     'REM'
              WHEN  rcode          IN ('0997', '0910', '0914')
              AND   rdesc      NOT LIKE 'POS:00003%'     THEN     NULL
              WHEN  rnumber_cnt     = 1               THEN     'RET'
           WHEN  top_rquantity     = '-1'               THEN     'RETXP'
           WHEN  top_rquantity     != '1'               THEN     NULL
           WHEN  seq_number     = 1               THEN     'RET'
                                     ELSE     'RETXP'
       END     AS transaction_type
  FROM     got_analytics
  ;
  

  I know that's not what you want, but I think that it is a step in the right direction.

• Helps the complex SQL update

  Hello
  
  I have a MyCounter table that is used to get the sequence numbers after N.
  For purposes of simplicity, let's just table MyCounter has 2 fields - maxNumber and currentNumber.
  maxNumber is the maximum sequence number that can be allocated.
  currentNumber is the last (current) sequence number.
  
  Assuming maxNumber is 1000, currentNumber is 500. If N is 10, the sequence should I get numbers: From = To = 510, 501
  
  I am new to Oracle. I had a first blow of dagger to write the SQL to update, but I am not convinced that it works.
  Can anyone give me suggestions please? Thank you...
  
  ---------
  
  DECLARE
  v_origValue NUMBER (10,0): = 0;
  v_allocated NUMBER (10,0): = 0;
  V_quantity NUMBER (10,0): = 0;
  
  BEGIN
  V_quantity: = 10; -assuming that 10 is a constant at this stage
  
  UPDATE MyCounter
  SET v_allocated = box
  When (currentNumber + v_quantity < = maxNumber) then restart
  else (maxNumber-currentNumber)
  end,
  v_origValue = (currentNumber),
  currentNumber = (currentNumber + v_allocated)
  WHERE the currentNumber < maxNumber;
  
  SELECT v_origValue + 1 as 'from', v_origValue + v_allocated 'to '.
  FROM DUAL;
  
  END;
  
  ---------
  
  Kind regards
  Judy

  Here is a small example of what you could do.

  SQL> create table t(type varchar2(10),max_val number, curr_val number)
    2  /
  
  Table created.
  
  SQL> insert into t values('type1',100,0)
    2  /
  
  1 row created.
  
  SQL> insert into t values('type2',1000,500)
    2  /
  
  1 row created.
  
  SQL> create or replace type my_seq as table of number
    2  /
  
  Type created.
  
  SQL> create or replace procedure get_sequence(ptype in varchar2, pcount in number, pseq out my_seq)
    2  as
    3     lcurr_val t.curr_val%type;
    4     lmax_val t.max_val%type;
    5     max_seq_exception exception;
    6  begin
    7     update t
    8        set curr_val = curr_val + pcount
    9      where type = ptype
   10     returning curr_val, max_val into lcurr_val, lmax_val;
   11
   12     if lcurr_val > lmax_val
   13     then
   14             rollback;
   15             raise max_seq_exception;
   16     else
   17             select lcurr_val - (level-1)
   18               bulk collect into pseq
   19               from dual
   20            connect by level <= pcount
   21              order by level desc;
   22
   23             commit;
   24     end if;
   25  exception
   26     when max_seq_exception then
   27             raise_application_error(-20001,'Max sequence exceeded');
   28  end;
   29  /
  
  Procedure created.
  
  SQL> declare
    2     lseq my_seq;
    3  begin
    4     get_sequence('type1',10,lseq);
    5     for i in 1..lseq.count
    6     loop
    7             dbms_output.put_line(lseq(i));
    8     end loop;
    9  end;
   10  /
  1
  2
  3
  4
  5
  6
  7
  8
  9
  10
  
  PL/SQL procedure successfully completed.
  
  SQL> select * from t
    2  /
  
  TYPE          MAX_VAL   CURR_VAL
  ---------- ---------- ----------
  type1             100         10
  type2            1000        500
  
  SQL> declare
    2     lseq my_seq;
    3  begin
    4     get_sequence('type1',100,lseq);
    5     for i in 1..lseq.count
    6     loop
    7             dbms_output.put_line(lseq(i));
    8     end loop;
    9  end;
   10  /
  declare
  *
  ERROR at line 1:
  ORA-20001: Max sequence exceeded
  ORA-06512: at "SYSADM.GET_SEQUENCE", line 27
  ORA-06512: at line 4
  

  But do not forget in a multiuser environment, there are many chances for you to get screwed!. Sequence is the best option in my opinion.

  Published by: Karthick_Arp on November 30, 2008 21:14

• Need help: complex SQL statement

  Select segment3, 0 "other income."
  , 0 "Sales Services".
  , 0 "staff costs".
  , 0 "other staff costs".
  , 0 'Deputy '.
  , 0 'trip '.
  , 0 "collaborator".
  , 0 'training '.
  , 0 'capital '.
  , nvl (round (sum (nvl(PERIOD_ACTIVITY_A,0) + nvl(PERIOD_ACTIVITY_E,0) + nvl(PRDOPEN_A,0) + nvl(PRDOPEN_E,0)), 0), 0) "Overhead"
  , 0 "financial expenses".
  of PAN. SNP_OP_DETAIL@OFA
  where segment3 = x.new_cc and period_name = p_period
  and segment2 in (select acc_t from the CRP_REPORT_ACCT_DETAIL_T where group_id = 10)
  Group of segment3
  
  Please, I do not understand select segment3, 0 'other income'... There is no column called 0, what is the 0s in the select statement?
  
  Secondly, I don't understand PAN. SNP_OP_DETAIL@OFA

  Means says:
  Select segment3, 0 "other income."
  , 0 "Sales Services".
  , 0 "staff costs".
  , 0 "other staff costs".
  , 0 'Deputy '.
  , 0 'trip '.
  , 0 "collaborator".
  , 0 'training '.
  , 0 'capital '.
  , nvl (round (sum (nvl(PERIOD_ACTIVITY_A,0) + nvl(PERIOD_ACTIVITY_E,0) + nvl(PRDOPEN_A,0) + nvl(PRDOPEN_E,0)), 0), 0) "Overhead"
  , 0 "financial expenses".
  of PAN. SNP_OP_DETAIL@OFA
  where segment3 = x.new_cc and period_name = p_period
  and segment2 in (select acc_t from the CRP_REPORT_ACCT_DETAIL_T where group_id = 10)
  Group of segment3

  Please, I do not understand select segment3, 0 'other income'... There is no column called 0, what is the 0s in the select statement?

  0 'other income' means to select a literal value of 0 (zero) with an alias of "other income". double quotes preserve the space and upper/lower case in the alias. This alias (in space) will be difficult to use because of the so called space later.

  Secondly, I don't understand PAN. SNP_OP_DETAIL@OFA

  PAN is probably a schema, SNP_OP_DETAIL a table or view and @OFA there is a link of database indicating the table in a different database

• TARGET RESULT COMPLEX SQL QUERY

  Hello
  
  I'm writing a sql, for the scenario below, but don't get it.
  
  There are two Table TEMP_AAA and TEMP_BBB,
  Under the table, we have values in descending order C_ID,
  
  Now to give a value in D_ID,.
  We must find the value of C_ID
  and if this C_ID value is contained in the B_ID,
  then we should display the value of allowance A_ID,
  
  The complexity here is,
  When you come down (in fact different levels) lines of C_ID,
  already consumed values should not be used.
  
  In this example,.
  for D_ID ABC says
  We have 1234 as C_ID and
  B_ID is 12345 which contains C_ID
  A_ID coresponding is 111 and 222.
  
  Next step for D_ID (same level): DEF since the C_ID is the same as the process described above
  We have A_ID 111 and 222.
  
  Then for D_ID GHI.
  We have 12 as C_ID and
  B_ID is 12345,12345,12045, which contains the C_ID
  A_ID coresponding allocation is 111,222,333,
  Now since 111 and 222 is already used in the ABC/DEF D_ID we would not have considered.
  A_ID would be 333
  
  Then for D_ID MNO.
  We have NULL as C_ID, which means all B_ID default matches (special account)
  B_ID is 12345,12345,12045,10045, which contains C_ID (default)
  A_ID coresponding allocation is 111,222,333,444
  Now given that 111,222,333 is already used in D_ID ABC/DEF/GHI/JKL we would not have considered.
  A_ID would be 444.
  
  
  
  Here is the Script and exit
  
  

   
  
  CREATE TABLE TEMP_AAA( A_ID VARCHAR2(10),  B_ID VARCHAR2(10));
  CREATE TABLE TEMP_BBB( C_ID VARCHAR2(10),  D_ID VARCHAR2(10));
  
  INSERT INTO TEMP_AAA(A_ID,B_ID) VALUES('111','12345');
  INSERT INTO TEMP_AAA(A_ID,B_ID) VALUES('222','12345');
  INSERT INTO TEMP_AAA(A_ID,B_ID) VALUES('333','12000');
  INSERT INTO TEMP_AAA(A_ID,B_ID) VALUES('444','10000');
  
  INSERT INTO TEMP_BBB(C_ID,D_ID) VALUES('1234','ABC');
  INSERT INTO TEMP_BBB(C_ID,D_ID) VALUES('1234','DEF');
  INSERT INTO TEMP_BBB(C_ID,D_ID) VALUES('12','GHI');
  INSERT INTO TEMP_BBB(C_ID,D_ID) VALUES('12','JKL');
  INSERT INTO TEMP_BBB(C_ID,D_ID) VALUES(NULL,'MNO');

  EXPECTED RESULTS
  ---------------
  

   
  
  D_ID      RESULT A_ID
  ---------------
  ABC     111
  ABC     222
  DEF     111
  DEF     222
  GHI     333
  JKL     333
  MNO     444

  Hello

  Thanks for posting the CREATE TABLE and INSERT statements; This helps a lot.

  Interesting problem!
  For each allocation a_id, you have to find the number of times where a c_id equaled and choose only the lines where it is the 1st time.
  This sounds like a job for the analytic RANK function:

  WITH     got_c_id_rank     AS
  (
       SELECT       b.d_id
       ,       a.a_id
       ,       RANK () OVER ( PARTITION BY  a_id
                                   ORDER BY      c_id     DESC     NULLS LAST
                        ) AS c_id_rank
       FROM       temp_aaa     a
       JOIN       temp_bbb     b     ON     a.b_id     LIKE b.c_id || '%'
  )
  SELECT       d_id
  ,       a_id          AS result_a_id
  FROM       got_c_id_rank
  WHERE       c_id_rank     = 1
  ORDER BY  d_id
  ,            result_a_id
  ;
  

• Problem SQL HELP during the installation of aviation the interactive dvd course!

  try to download interactive dvd Aviation course and be sktc kingschool SQL error does not exist... Check the State of SQL express instance and I install the classes I get a SsRestart occurred must close and send error report... Any help would be great!    Help!

  I suggest that you contact the manufacturer of the interactive DVD software to help get the program installed.

  They would be better able to help with your problem.

  Thank you

  Marilyn

• SQL Help: Add flag for several tables

  Hello

  I have an existing SQL as follows:

  Select the storeId field,

  (select case when count (1) > 0 0 otherwise then 1 end)

  of tb_bbb

  where tb_bbb.id = tb_base.baseId) AS conInd

  of tb_base;

  I would now add the table tb_ccc for the value of conInd. I try the following SQL:

  Select the storeId field,

  (select case when count (1) > 0 0 otherwise then 1 end)

  Of

  (select tb_bbb.id where tb_bbb.id = tb_base.baseId)

  Union

  Select tb_ccc.id where tb_ccc.id = tb_base.basdId) AS conInd

  of tb_base;

  However, it does not work. Any idea on this SQL?  Thanks for your help!

  Hello

  Here is another way, which may be more effective because it uses EXISTS (which can quit smoking as soon as it finds a match) instead of COUNT (who continues to go, so he can tell you if there are 999 or 1000 matches, even if you don't care to).

  SELECT storeid

  CASE

  WHEN THERE IS)

  SELECT 1

  OF tb_bbb

  WHERE bbb_id = m.baseid

  ) 1 THEN

  WHEN THERE IS)

  SELECT 1

  OF tb_ccc

  WHERE ccc_id = m.baseid

  ) 1 THEN

  0 OTHERWISE

  END AS conlnd

  OF tb_base m

  ;

• Updated SQL help

  Hi all

  I need correct column value serNo to the right as a result. It suppose to be the sequence number for each storeID. Any suggestion on the SQL?

  create table test1 (storeID number (5), number serNo (5));

  insert into table test1 values (100, 1);

  insert into table test1 values (200, 1);

  insert into table test1 values (100, 1);

  insert into table test1 values (200, 1);

  insert into table test1 values (100, 1);

  insert into table test1 values (100, 1);

  I want to have the following results in test1 after conversion:

  1 100

  2 100

  3 100

  4 100

  1 200

  2 200

  Thank you

  Hello

  942572 wrote:

  Hi all

  I need correct column value serNo to the right as a result. It suppose to be the sequence number for each storeID. Any suggestion on the SQL?

  create table test1 (storeID number (5), number serNo (5));

  insert into table test1 values (100, 1);

  insert into table test1 values (200, 1);

  insert into table test1 values (100, 1);

  insert into table test1 values (200, 1);

  insert into table test1 values (100, 1);

  insert into table test1 values (100, 1);

  I want to have the following results in test1 after conversion:

  1 100

  2 100

  3 100

  4 100

  1 200

  2 200

  Thank you

  Use the ROW_NUMBER analytic function to generate the new sernos, like this:

  MERGE INTO dst test1

  WITH THE HELP OF)

  SELECT ROWID AS r_id

  ROW_NUMBER () OVER (PARTITION BY storeid

  ORDER BY NULL

  ) AS serno

  OF test1

  )               src

  ON (summer time. ROWID = src.r_id)

  WHEN MATCHED THEN UPDATE

  SET dst.serno = src.serno

  ;

  Really, your table must have a primary key.  Because it's not working, I had to ROWID allows you to uniquely identify the lines.

• SQL Help: How to collect summary number in group by?

  
  Hello

  I have the following table:

  create table tb_class_info (classNbr number (5), number (2) of ClassType, classTeacherNbr number (4));

  insert into tb_class_info values (101, 1, 12);

  insert into tb_class_info values (001, 2, 12);

  insert into tb_class_info_values (001, 2, 13);

  insert into tb_class_info_values (002, 2, 12);

  insert into tb_class_info_values (002, 2, 12);

  I would like to get statistics on the cnt, cnt for classType = 2 classTeacherNbr classNbr. I expect to have the following results:

  classNbr classType2Cnt classTeacherCnt

  101                0                         0

  001                2                          2

  002                2                          1

  The following SQL code I use:

  SELECT classNbr,

  SUM (CASE when classType = 2 THEN 1 ELSE END 0) as classTypeCnt,

  SUM (CASE when classType = 2 SO unique classTeacherNbr ELSE 0 END) as classTeacherCnt

  OF tb_class_info

  GROUP BY classNbr;

  However, I have the ' ORA-00905: lack of keyword ' error. Any suggestion on this subject?

  Thanks for your help!

  Select classnbr

  sum (case when classtype = 2 then 1 end)

  , count (distinct case when classtype = 2 then teachernbr end)

  of tb_class_into

  Classnbr group;

• PL/SQL help

  I'm pretty new in pl/sql, I want to update a column with the next number.

  Here's an example-

  Table: political

  Policy_no	Sequence	Company_name
  101	0	ABC PTY LTD.
  101	0	XYZ PTY LTD
  103	0	Test PTY LTD
  104	0	Test101 PTY LTD
  101	0	Pearl PTY LTD
  104	0	MyZore pty ltd

  I would get the date in the following format

  After Update - table

  Policy_no	Sequence	Company_name
  101	1	ABC PTY LTD.
  101	2	XYZ PTY LTD
  101	3	Pearl PTY LTD
  103	1	Test PTY LTD
  104	1	Test101 PTY LTD
  104	2	MyZore pty ltd

  Basically, I want to update the column sequence based on the Policy_No group.

  Can someone please help.

  >

  Basically, I want to update the column sequence based on the Policy_No group.

  >

  This query will show you how to generate the values you need.

  >

  Select empno, deptno, row_number() over (partition by deptno) myRow arrested by deptno from EMP

  EMPNO, DEPTNO, MYROW

  7782,10,1

  7839,10,2

  7934,10,3

  7566,20,1

  7902,20,2

  7876,20,3

  7369,20,4

  7788,20,5

  7521,30,1

  7844,30,2

  7499,30,3

  7900,30,4

  7698,30,5

  7654,30,6

  >

  According to the number of rows in the table, you might be better to create a new table using DEC or a query like this.

• Multiple Table join without loss of data - SQL help

  Hello
  I have the following tables 3 employee as indicated in the scripts below, please help in writing SQL that connects all these 3 tables and displays the output as shown below
  
  Required output
  EMP ID Loc Value 1 Value 2 Value 3                    
  1     A     3     1     6
  1 B 4 9 null
  Value null null 1 Wh 8
  2 Z 4 9 null
  3 R 0 null 1
  4 Y 1 null null
  5 O 7 7 null
  Value null null 5 Wh 5
  6 M 7 null 7
  6 B null null 8
  
  / * Create Tables * /.
  Create table Emp1 (Empid number, Loc Varchar2 (10), Value_1 number);
  Create the Emp2 Table (Empid number, Loc Varchar2 (10), Value_2 number);
  Create table emp3 (empid number, loc varchar2 (10), Value_3);
  
  / * Insert Scripts * /.
  Insert Into Emp1 Values (1, 'A', 3);
  Insert Into Emp1 Values (2, 'Z', 4);
  Insert Into Emp1 Values (3, 'R', 0);
  Insert Into Emp1 Values (4, 'Y', 1);
  Insert Into Emp1 Values (5, 'o', 7);
  Insert Into Emp1 Values (6, am', 7);
  
  Insert Into Emp2 Values (1, 'A', 1);
  Insert Into Emp2 Values (1, 'B', 4);
  Insert Into Emp2 Values (1, 'Wh', 8);
  Insert Into Emp2 Values (2, 'Z', 9);
  Insert Into Emp2 Values (5, 'o', 7);
  Insert Into Emp2 Values (5, 'Wh', 5);
  
  Insert Into Emp3 Values (1, 'A', 6);
  Insert Into Emp3 Values (1, 'B', 9);
  Insert Into Emp3 Values (3, 'R', 1);
  Insert Into Emp3 Values (6, am', 7);
  Insert Into Emp3 Values (6, 'B', 8);
  Commit;
  
  
  Thank you
  Varun

  Like this?

  SQL> select nvl(e.empid, e3.empid) empid
    2       , nvl(e.loc, e3.loc) loc
    3       , e.value_1
    4       , e.value_2
    5       , e3.value_3
    6    from (
    7            select nvl(e1.empid, e2.empid) empid
    8                 , nvl(e1.loc, e2.loc) loc
    9                 , e1.value_1
   10                 , e2.value_2
   11              from emp1 e1
   12              full join emp2 e2
   13                on e1.empid = e2.empid
   14               and e1.loc = e2.loc
   15         ) e
   16    full join emp3 e3
   17      on e.empid = e3.empid
   18     and e.loc = e3.loc
   19   order
   20      by empid
   21       , loc
   22  /
  
       EMPID LOC           VALUE_1    VALUE_2    VALUE_3
  ---------- ---------- ---------- ---------- ----------
           1 A                   3          1          6
           1 B                              4          9
           1 Wh                             8
           2 Z                   4          9
           3 R                   0                     1
           4 Y                   1
           5 O                   7          7
           5 Wh                             5
           6 B                                         8
           6 M                   7                     7
  
  10 rows selected.
  
  SQL> 
  

• SQL help. Identify changes to a field.

  Greetings!
  PS/SQL is not an option for me. I need help to use SQL, if possible for the following scenario.
  Oracle 10G.
  
  Table: JOB_DATA
  
  EMPLID, DATE_EFF, DEPTID, JOBCODE
  100, 01/11/2012, 34567, MNG
  100, 01/10/2012, 34567, SUP
  100, 01/09/2012, 28967, MNG
  100, 15/08/2012, 28967, SUP
  100,6/30/2012,15879, MNG
  
  I need to get the following records only, in other words, every time that changes in the Department ID.
  
  100, 01/10/2012, 34567, SUP
  100, 15/08/2012, 28967, SUP
  100,6/30/2012,15879, MNG
  
  Thanks in advance.

  It looks like you want something like

  SELECT *
    FROM (SELECT j.*, lag(deptid) over (partition by emplid order by date_eff) prior_deptid
            FROM job_data j)
   WHERE prior_deptid IS NULL
      OR prior_deptid != deptid
  

  Justin

• SQL help on taking in SHORT

  Hello
  
  I have the following table
  

  EmployeeID | ProjectID | Tasks
  EMP 01     | PID 01    | 5
  EMP 01     | PID 01    | 1
  EMP 01     | PID 02    | 8
  EMP 01     | PID 03    | 9
  EMP 02     | PID 01    | 7
  EMP 02     | PID 02    | 4
  EMP 02     | PID 02    | 3
  EMP 03     | PID 04    | 1
  EMP 04     | PID 04    | 8
  EMP 04     | PID 04    | 7

  I need to write the sql query to get the following results
  

  EmployeeID | ProjectID | EmpTasks | ProjectTasks
  EMP 01     | PID 01    | 23       | 6
  EMP 01     | PID 01    | 23       | 6
  EMP 01     | PID 02    | 23       | 8
  EMP 01     | PID 03    | 23       | 9
  EMP 02     | PID 01    | 14       | 7
  EMP 02     | PID 02    | 14       | 7
  EMP 02     | PID 02    | 14       | 7
  EMP 03     | PID 04    | 1        | 1
  EMP 04     | PID 04    | 15       | 15
  EMP 04     | PID 04    | 15       | 15

  Any help?
  
  Thank you

  This should do it:

  WITH t AS
  (SELECT 'EMP 01' EmployeeID, 'PID 01' ProjectID, 5 Tasks FROM dual UNION ALL
   SELECT 'EMP 01' EmployeeID, 'PID 01' ProjectID, 1 Tasks FROM dual UNION ALL
   SELECT 'EMP 01' EmployeeID, 'PID 02' ProjectID, 8 Tasks FROM dual UNION ALL
   SELECT 'EMP 01' EmployeeID, 'PID 03' ProjectID, 9 Tasks FROM dual UNION ALL
   SELECT 'EMP 02' EmployeeID, 'PID 01' ProjectID, 7 Tasks FROM dual UNION ALL
   SELECT 'EMP 02' EmployeeID, 'PID 02' ProjectID, 4 Tasks FROM dual UNION ALL
   SELECT 'EMP 02' EmployeeID, 'PID 02' ProjectID, 3 Tasks FROM dual UNION ALL
   SELECT 'EMP 03' EmployeeID, 'PID 04' ProjectID, 1 Tasks FROM dual UNION ALL
   SELECT 'EMP 04' EmployeeID, 'PID 04' ProjectID, 8 Tasks FROM dual UNION ALL
   SELECT 'EMP 04' EmployeeID, 'PID 04' ProjectID, 7 Tasks FROM dual)
  SELECT EmployeeID,
         ProjectID,
         SUM(Tasks) OVER (PARTITION BY EmployeeID) EmpTasks,
         SUM(Tasks) OVER (PARTITION BY EmployeeID, ProjectID) ProjectTasks
  FROM   t
  

• REGEXP_LIKE sql help...

  I have a first_name column in my table of students with the following characteristics:
  
  REDA
  Kris
  Tanvi
  Shweta
  Rajendra
  Kapil
  Vipin
  Sandeep
  
  Now I'm trying to use the regexp_like function to find the names that begin the letter s and end with the letter p, but I am not able to achieve that I provided here my request please help:
  
  Select first_name
  the student
  where regexp_like (name, ' ^ sp$ ');
  
  Concerning
  Rahul

  Mac_Freak_Rahul wrote:
  where regexp_like (name, ' ^ sp$ ');

  Describe your regular expression is to find the string "sp". You do not include buildings for characters that could go p and between s. Here is an example that may help:

  SQL> WITH student AS
    2  (
    3          SELECT 'raj' AS first_name FROM dual UNION ALL
    4          SELECT 'kris' AS first_name FROM dual UNION ALL
    5          SELECT 'tanvi' AS first_name FROM dual UNION ALL
    6          SELECT 'shweta' AS first_name FROM dual UNION ALL
    7          SELECT 'rajendra' AS first_name FROM dual UNION ALL
    8          SELECT 'kapil' AS first_name FROM dual UNION ALL
    9          SELECT 'vipin' AS first_name FROM dual UNION ALL
   10          SELECT 'sandeep' AS first_name FROM dual
   11  )
   12  SELECT first_name
   13  FROM   student
   14  WHERE  REGEXP_LIKE(first_name,'^s.*p$')
   15  ;
  
  FIRST_NA
  --------
  sandeep
  

• Coding SQL help

  Hi all
  I need help on getting the sql code.
  Here is how the example data are arranged in my data table called for example.
  
  favorite colors of name age
  John doe 15 Red: Blue: Green
  
  Notice how the data in the column "favorite colors", defined the values selected from a list of several.
  I need the sql code so that it can display like this.
  
  favorite colors of name age
  John doe 15 Red
  John doe 15 blue
  John doe 15 green
  
  Thank you

  Hi bigmac704,

  There may be other ways to do it, but one way is to use apex_util.string_to_table with a function in the pipeline:

  (1) create a type to define the line of your favorite color of the person:

  CREATE OR REPLACE TYPE  PERSON_COLOR as object (name varchar2(30), age number(3, 0), fav_color varchar2(10));
  

  (2) create a type to define a "picture" of your person-favorite color lines:

  CREATE OR REPLACE TYPE  PERSON_COLOR_TBL AS TABLE OF person_color
  

  (3) create the following function in the pipeline:

  CREATE OR REPLACE FUNCTION get_person_colors RETURN person_color_tbl
  PIPELINED IS
    out_rec    person_color := person_color(null, null, null);
    l_vc_arr2  APEX_APPLICATION_GLOBAL.VC_ARR2;
  BEGIN
    FOR r IN (SELECT * FROM person_colors) LOOP
      out_rec.name := r.name;
      out_rec.age := r.age;
      l_vc_arr2 := APEX_UTIL.STRING_TO_TABLE(r.fav_colors);
      // for each color, pipe a row
      FOR i IN 1..l_vc_arr2.count LOOP
        out_rec.fav_color := l_vc_arr2(i);
        pipe row(out_rec);
      END LOOP;
    END LOOP;
    RETURN;
  END;
  

  Now you can use this function in the pipeline, just as you would use a table:

  SELECT * FROM TABLE(get_person_colors);
  NAME AGE FAV_COLOR
  john doe 15 red
  john doe 15 blue
  john doe 15 green 
  
  SELECT * FROM TABLE(get_person_colors) WHERE name = 'john doe';
  NAME AGE FAV_COLOR
  john doe 15 red
  john doe 15 blue
  john doe 15 green 
  
  SELECT * FROM TABLE(get_person_colors) WHERE name = 'jim smith';
  no data found
  

  Hope this helps,
  John

  If you find this information useful, please do not forget to mark the 'useful' or 'correct' post so that others benefit as well.