Lag function

Oracle version: 11.0.2.0.2.0

Dear all

I have a question where I am using tables Tab1 and Tab2

Struncture of tab1 is

Store_no NUMBER,

Sell_date DATE

Sell quantity number,

number of item_no

to tab2

Number of Item_no

date date1,

number of item_price

I need know the total price of the article sold for 2 days (current as previous)

Select the store, sell_date, item_no, sell_qty * item_price, lag (sell_qty * item_price, 1) on the partition by (store_no, item_no order of sell_date)

of tab1, tab2

where tab1.item_no = tab2.item_no

and sell_date = date1

Question1) tables have millions of lines... The lag function is causing the query to run for hours...

Could you please suggest if we capture these sorts of queries where Lag/lead are used.

Your query does not much sense. Analytical functions are applied to already recovered lines, and since you have sell_date = date1, all you get is one day sales. All you need is:

Select the store

Max (sell_date),

tab1.item_no,

sum (case sell_date when date1 then sell_qty * end item_price) day_sales_amount,.

sum (case sell_date when date1 - sell_qty then 1 * item_price end) prev_day_sales_amount

of tab1.

tab2

where tab1.item_no = tab2.item_no

and sell_date in the fields (date1, date1 - 1).

Store group,

tab1.item_no;

SY.

Tags: Database

Similar Questions

  • In Oracle lag function

    Hi all

    I want to understand the inner workings of the Lag function in Oracle, which could help to solve my problem.

    I have the table as the structure below.

    P_ID ENTRY_TIMESTAMP SIZE C_ID
    152871 NOVEMBER 15 05.45.00 PM3092
    1529506.00.00 PM 1ST NOVEMBER 151592
    1530606.00.00 PM 1ST NOVEMBER 15192

    Now, when I write a query as below:

    Select the size, CURR_TIMESTAMP TO_TIMESTAMP (TO_CHAR (ENTRY_TIMESTAMP, 'YYYYMMDDHH24MI'), 'YYYYMMDDHH24MI'),

    TO_TIMESTAMP (TO_CHAR (LAG (ENTRY_TIMESTAMP) OVER (PARTITION BY ORDER BY ENTRY_TIMESTAMP ASC C_ID), 'YYYYMMDDHH24MI'), 'YYYYMMDDHH24MI') LAG_TIMESTAMP

    OF Entry_log

    where ENTRY_TIMESTAMP BETWEEN to_date (1 November 15 18.00.00','DD-MON-YY HH24.MI.) SS') AND to_date (1 November 15 18.00.00','DD-MON-YY HH24.MI.) SS')

    Query will lead to give Lag_timestamp as 06.00.00 PM for size = 1 but will be NULL for size = 15.

    Why is that Oracle will never lag_timestamp = 06.00.00 PM for size = 15.

    Could someone please explain this behavior.

    I'm using Oracle 11 g.

    Thank you

    I had a small mistake in the above query. now, I fixed it

  • LAG function alternative queries

    Dear all,

    WITH t AS

    (SELECT 1 AS num, "test1" AS status

    OF THE DOUBLE

    UNION ALL

    SELECT 2 AS num, "test2" LIKE status

    OF THE DOUBLE

    UNION ALL

    SELECT 3 AS num, "test3" LIKE status

    OF THE DOUBLE

    UNION ALL

    SELECT the 4 ACEs num, "test4" LIKE status

    OF THE DOUBLE

    UNION ALL

    SELECT 5 AS num, "test5" LIKE status

    THE DOUBLE)

    SELECT the State, LAG (status 1) OVER (ORDER BY num) AS previous_status

    T

    Please help me to write search above without LAG function like oracle forms10g unsupported feature LAG.

    Hello, you can join the table itself for the results. This will work only if you have a number in ascending order.

    WITH t AS

    (SELECT 1 AS num, "test1" AS status

    OF THE DOUBLE

    UNION ALL

    SELECT 2 AS num, "test2" LIKE status

    OF THE DOUBLE

    UNION ALL

    SELECT 3 AS num, "test3" LIKE status

    OF THE DOUBLE

    UNION ALL

    SELECT the 4 ACEs num, "test4" LIKE status

    OF THE DOUBLE

    UNION ALL

    SELECT 5 AS num, "test5" LIKE status

    THE DOUBLE)

    T1.status, t2.status SELECT previous_status

    T t1, t t2

    where t2.num (+) = t1.num - 1

    order of t1.status

    Best regards, Andrei

  • LAG function to get the value of the measurement for parent

    Hello

    I'm interested if there is no function in OLAP DML, which is similar to the function of TROLLING. LAG function returns the measured value for the previous dimension member (or a few members of the back dimension). I would like to know if this function exists which does not return a meaningless measurement for the dimension member value, but the measured value for the parent of the live dimension member.

    Thanks in advance

    Peter

    _parentrel relationship identifies the parent of each Member of the dimension of a hierarchy.

    use:
    the time - dimension
    time_parentrel - Member parent expression
    time_parentrel (time_hierlist 'CALENDAR') - expression for parent along the CALENDAR hierarchy

    Sales - measure
    Sales (Time time_parentrel) - expression returning sales for the parent of the current Member along the existing hierarchy.
    sales (time time_parentrel (time_hierlist 'EXERCISE'))-expression returning the sales for the parent of the current Member along the FISCAL hierarchy.

    You can search the user's Guide the OLAP DML for many examples of the use of the OLAP DML commands/functions/objects in expressions.
    All of these objects are automatically created/maintained while retaining the dimension from the front end (MN).

    For example:
    (11g) OLAP DML users guide notions-> about-> OLAP DML objects given OLAP DML-> objects that support the use of hierarchical Dimensions section.

  • lag function, part two

    Oracle 11.2.0.1.0 SE - one on OL 5.6 x 86-64

    Still faced with a LAG.

    Given a table like this
    create table emp_hist (
     sample_date  date,
     lname   varchar2(10),
     dept    varchar2(10),
     sal     number
     )
;
    Table is added every day with
    insert into emp_hist (select sysdate,
                       lname,
                       dept,
                       sal
                from emp)
;
    Of course, most of the time, most of the employees will have no change from the previous. What I'm trying to do is filtering all the records of 'no change' and get off at only changes... amount of change and when it appeared

    Maybe this:
    dept         lname        sample_date     sal
----         -----        -----------     -------
10           smith        1-Aug-2012      10000
                          2-Aug-2012      20000

10           jones        5-Aug-2012      50000
                          6-Aug-2012      60000

20           snuffy       4-Aug-2012      30000
                          5-Aug-2012      40000
    or (better)
    dept         lname        change_date     prev_sal  curr_sal
----         -----        -----------     -------   --------
10           smith        2-Aug-2012      10000     20000 
10           jones        6-Aug-2012      50000     60000
20           snuffy       5-Aug-2012      30000     40000
    Wish I could show what I tried so far, but I'm so stuck, I can't even that.

    Hello

    Looks like:

    WITH     got_prev_sal     AS
(
     SELECT       dept
     ,       lname
     ,       sample_date          AS change_date
     ,       LAG (sal) OVER ( PARTITION BY  lname
                                 ORDER BY        sample_date
                     )      AS prev_sal
     ,       sal                AS curr_sal
     FROM       emp_hist
)
SELECT       *
FROM       got_prev_sal
WHERE       prev_sal     != curr_sal
ORDER BY  lname
,            change_date
;

    In any application data (or subquery), analytical functions are calculated after the clause WHERE this request has been applied. If you need to use the results of an analytic function in a WHERE clause, calculate the analytical function in a subquery (such as got_prev_sal, above). You can then use the value returned in the WHERE clause (or elsewhere) in the main query.

    I hope that answers your question.
    If not, post a small example data (instructions INSERT, to go with the CREATE TABLE statement, you have already published) and also publish outcomes from these data.
    Point where the above query is to produce erroneous results, and explain, using specific examples, how you get the right results in these places.
    See the FAQ forum {message identifier: = 9360002}

    Published by: Frank Kulash, August 7, 2012 16:29

  • get my head wrapped around the LAG function

    Oracle 11.2.0.4 Std Ed

    drop table change_hist;

    create table change_hist (file_name varchar2 (30),)

    date of sample_date,

    number of blocks

    );

    insert into change_hist values (to_date('2015-07-01','yyyy-mm-dd'), 'file1', 5);

    insert into change_hist values (to_date('2015-07-02','yyyy-mm-dd'), 'file1', 5);

    insert into change_hist values (to_date('2015-07-03','yyyy-mm-dd'), 'file1', 6);

    insert into change_hist values (to_date('2015-07-04','yyyy-mm-dd'), 'file1', 6);

    insert into change_hist values (to_date('2015-07-05','yyyy-mm-dd'), 'file1', 6);

    insert into change_hist values (to_date('2015-07-01','yyyy-mm-dd'), 'file2', 5);

    insert into change_hist values (to_date('2015-07-02','yyyy-mm-dd'), 'file2', 5);

    insert into change_hist values (to_date('2015-07-03','yyyy-mm-dd'), 'file2', 5);

    insert into change_hist values (to_date('2015-07-04','yyyy-mm-dd'), 'file2', 5);

    insert into change_hist values (to_date('2015-07-05','yyyy-mm-dd'), 'file2', 5);

    insert into change_hist values ('file3', to_date('2015-07-01','yyyy-mm-dd'), 5);

    insert into change_hist values ('file3', to_date('2015-07-02','yyyy-mm-dd'), 6);

    insert into change_hist values ('file3', to_date('2015-07-03','yyyy-mm-dd'), 6);

    insert into change_hist values ('file3', to_date('2015-07-04','yyyy-mm-dd'), 6);

    insert into change_hist values ('file3', to_date('2015-07-05','yyyy-mm-dd'), 6);

    Desired result

    file_name sample_date prev_value curr_value

    File1 2015-07-03 5 6

    File2 2015-07-02 5 6

    My first attempt...

    SQL > select file_name,.

    2 sample_date,

    3 lag (blocks, 1) (order sample_date) as prev_blocks

    4 change_hist;

    FILE_NAME SAMPLE_DA PREV_BLOCKS

    ------------------------------ --------- -----------

    July 1, 15 file1

    July 1, 15 5 file3

    July 1, 15 5 File2

    File2 2 July 15 5

    File1 2 July 15 5

    file3 2 July 15 5

    File1 July 3, 15 6

    File2 July 3, 15 6

    file3 July 3, 15 5

    File1 4 July 15 6

    file3 4 July 15 6

    FILE_NAME SAMPLE_DA PREV_BLOCKS

    ------------------------------ --------- -----------

    File2 4 July 15 6

    File2 July 5 15 5

    File1 July 5 15 5

    file3 5 July 15 6

    15 selected lines.

    SQL >

    SQL > spo off

      1  select * from (select file_name, sample_date, lag(blocks) over (partition by file_name order by sample_date) prev,
blocks
  2  from change_hist)
  3* where prev != blocks
SQL> /                                                                                                             

FILE_NAME                      SAMPLE_DA      PREV    BLOCKS
------------------------------ --------- ---------- ----------
file1                          03-JUL-15          5          6
file3                          02-JUL-15          5          6

    You know... you need to partition the file name

    Edit: in the interest of beat some people with a force equal regardless of their status, you didn't really know we tell how to achieve these results, so I could refer to the FAQ on how to ask a question

  • Grouping and the LAG function

    Oracle 10.2.0.4 SE - One on OL 5 x 86-64

    Trying to write reports on statspack tables. In view of this query
    select     
--     s1.ucomment
     w1.event
,     s1.snap_id
,     to_char(s1.snap_time,'dd-Mon-yyyy hh24:mi:ss') snap_time
,     lag(w1.total_waits)
over (order by s1.snap_id) prev_val,
      w1.total_waits - lag(w1.total_waits)
                        over (order by s1.snap_id) delta_val
,     lag(w1.time_waited_micro)
over (order by s1.snap_id) prev_val,
      w1.time_waited_micro - lag(w1.time_waited_micro)
                        over (order by s1.snap_id) delta_val
from      stats$snapshot s1
,     stats$system_event w1
where s1.snap_id = w1.snap_id
  and w1.event = 'db file sequential read'
order by w1.event
,     s1.snap_id
;
EVENT                                                               SNAP_ID SNAP_TIME              PREV_VAL  DELTA_VAL   PREV_VAL  DELTA_VAL
---------------------------------------------------------------- ---------- -------------------- ---------- ---------- ---------- ----------
db file sequential read                                                   1 25-Apr-2013 11:00:51
db file sequential read                                                   2 25-Apr-2013 12:00:51     959235      39179  949588625   29088092
db file sequential read                                                   3 25-Apr-2013 13:00:51     998414      29632  978676717   24938574
db file sequential read                                                   4 25-Apr-2013 14:00:52    1028046      21547 1003615291   23101379

4 rows selected.
    Works perfectly. But I would like to delete the ' where... w1.event = 'db file sequential read' and get a report on all events. Of course, simply commenting on this condition means that the first occurrence of an event is "shifted" against the last occurrence of the previous event. Can not find a way to get the logic that do break and start the 'delay' all over again.

    You use lag (...) over (partition by the order of the events of snap_id), i.e. "partition by the event" will do lag a ' star on "for each different event

  • Row1-row2(without lead and lag function)

    Hello

    I wanted less row1 - row2 can I could use at least conditions he same for 1000 lines and should not use the features of lead and lag


    Thank you

    It if self explaining IMHO

    with data as (
select 9 r from dual union all
select 7 from dual union all
select 6 from dual union all
select 3 from dual
)
, rdata as (
select r, row_number() over (order by r desc) rn
from data
)

select
d1.r - nvl(d2.r, '0') res
from
rdata d1, rdata d2
where
d1.rn +1 =d2.rn(+)

RES
2
1
3
3

    Concerning

  • Using the LAG function to find a previous record

    Hey,.

    I tried searching the forum for this, but I didn't actually even know what to look for, so I create a new thread.

    The SQL below displays a list of prices:
    WITH T AS (
  SELECT 1  AS ID, 'PERM' AS TYPE, 100 AS PRICE, SYSDATE + 1 AS START_DATE FROM DUAL UNION
  SELECT 3  AS ID, 'TEMP' AS TYPE, 90  AS PRICE, SYSDATE + 2 AS START_DATE FROM DUAL UNION
  SELECT 7  AS ID, 'TEMP' AS TYPE, 80  AS PRICE, SYSDATE + 3 AS START_DATE FROM DUAL UNION
  SELECT 8  AS ID, 'PERM' AS TYPE, 75  AS PRICE, SYSDATE + 4 AS START_DATE FROM DUAL UNION
  SELECT 16 AS ID, 'TEMP' AS TYPE, 70  AS PRICE, SYSDATE + 5 AS START_DATE FROM DUAL UNION
  SELECT 20 AS ID, 'TEMP' AS TYPE, 60  AS PRICE, SYSDATE + 6 AS START_DATE FROM DUAL UNION
  SELECT 34 AS ID, 'TEMP' AS TYPE, 50  AS PRICE, SYSDATE + 7 AS START_DATE FROM DUAL
)
  SELECT T.ID
       , T.TYPE
       , T.PRICE
       , TRUNC (T.START_DATE) AS START_DATE
       , CASE 
           WHEN T.TYPE = 'PERM' 
             THEN T.ID 
             ELSE LAG (T.ID, 1, NULL) OVER (PARTITION BY NULL ORDER BY T.ID)
         END AS BASE_ID
    FROM T
ORDER BY T.START_DATE
    The challenge is to produce this output:
    ID TYPE PRICE BASE_ID
 1 PERM   100       1
 3 TEMP    90       1
 7 TEMP    80       1
 8 PERM    75       8
16 TEMP    70       8
20 TEMP    60       8
34 TEMP    50       8
    What I want to achieve is to bring a column with the ID of the most recent prices of TEMP PERM price.
    and it's own ID of prices of PERM.

    My attempt using LAG to browse the set of records, but it uses 1 statically. If there was a way to arrive at a number
    for each TEMP price say to what extent is the most recent Perm, then I could use that number instead of 1.

    Something like:
    ID TYPE PRICE DISTANCE_FROM_PREV_PERM
 1 PERM   100                       0
 3 TEMP    90                       1
 7 TEMP    80                       2
 8 PERM    75                       0
16 TEMP    70                       1
20 TEMP    60                       2
34 TEMP    50                       3
    Any help will be greatly appreciated.

    Thank you.

    Maybe

    select id,type,price,
       last_value(base_id) ignore nulls over (order by the_row) base_id
/*     last_value(base_id ignore nulls) over (order by the_row) base_id  -- old way */
  from (select id,type,price,start_date,
               case type when 'PERM' then id end base_id,
               row_number() over (order by start_date) the_row
          from t
       )

    Concerning

    Etbin

  • I think I need a LEAD/LAG function?

    Hello

    Oracle 10.2.0.4

    I have a table that contains an atrribute used over time.  Number of lines does not change the value and is there only to maintain integrity between the lines of other tables.

    CREATE THE EMP_ATTR TABLE (EMPID VARCHAR2 (8) NOT NULL, ASOFDATE DATE NOT NULL, ATTR VARCHAR2 (13));

    Values inserted INTO the EMP_ATTR ("'001', January 1, 2010 ', ' XX');

    INSERT INTO EMP_ATTR VALUES ('001', FEBRUARY 2, 2011 "," YY");

    INSERT INTO EMP_ATTR VALUES ('001', MARCH 4, 2011 "," YY");

    INSERT INTO EMP_ATTR VALUES ('002', JANUARY 1, 2010 "," TT");

    INSERT INTO EMP_ATTR VALUES (' '002', 1ST FEBRUARY 2010 TT ', ' ');

    INSERT INTO EMP_ATTR VALUES (' '002', 1 DECEMBER 2013 ', ' TT');

    INSERT INTO EMP_ATTR VALUES (' '003', JANUARY 1, 2010 ', ' EW');

    INSERT INTO EMP_ATTR VALUES (' '003', MARCH 2, 2010 ', ' RT');

    INSERT INTO EMP_ATTR VALUES (' '003', 1 NOVEMBER 2013 EW ', ' ');

    INSERT INTO EMP_ATTR VALUES (' '003', 1 DECEMBER 2013 EW ', ' ');

    What I have to do is to analyze the table, and for each return line when the person assigned this attribute, without going through another attribute assigned in the meantime

    So I need something that will return

    EMPID, ATTR, ASOFDATE, FIRST_ASOFDATE

    001, XX, JANUARY 1, 2010, JANUARY 1, 2010

    001, YY, FEBRUARY 2, 2011, FEBRUARY 2, 2011

    001, YY, March 4, 2011, February 2, 2011 <-as this was awarded in February and the line of March has not changed the attribute

    002, TT, JANUARY 1, 2010, JANUARY 1, 2010

    002, TT, FEBRUARY 1, 2010, JANUARY 1, 2010

    002, TT, DECEMBER 1, 2013, JANUARY 1, 2010

    003, EW, JANUARY 1, 2010, JANUARY 1, 2010

    003, RT, MARCH 2, 2010, MARCH 2, 2010

    003, EW, November 1, 2013, November 1, 2013 <-EW was previously assigned in 2010, but they have a RT line between Mar-2010 so it's seen as a reboot

    003, EW, December 1, 2013, November 1, 2013 <-there was no intermediate attribute inherit so assignment NOV forward

    Can be

    SQL > select empid, asofdate, attr, asofdate_first
    25 years of emp_attr
    model 26
    27 partition by (empid)
    size 28 per (row_number() over (partition by asofdate order empid) rno)
    29 measures (asofdate, attr, sysdate asofdate_first, ' ' attr_prev)
    30 rules
    (31)
    32 attr_prev [any] = attr [cv () - 1],
    33 asofdate_first [any] = decode (attr_prev [cv ()], attr [cv ()], [cv () - 1] asofdate_first, asofdate [cv ()])
    34)
    35 order by empid, asofdate;

    EMP ASOFDATE TO ASOFDATE_
    --- --------- -- ---------
    001-1 JANUARY 10 XX 1 JANUARY 10
    001 2 FEBRUARY 11 AA 2 FEBRUARY 11
    001 4 MARCH 11 AA 2 FEBRUARY 11
    002 1 JANUARY 10 TT 1 JANUARY 10
    002 1 FEBRUARY 10 TT 1 JANUARY 10
    002 1 DECEMBER 13 TT 1 JANUARY 10
    003 1 JANUARY 10 EW 1 JANUARY 10
    003 2 MARCH 10 RT 2 MARCH 10
    1ST NOVEMBER 13 003 EW 1 NOVEMBER 13
    1ST DECEMBER 13 003 EW 1 NOVEMBER 13

    10 selected lines.

  • [8i] need help prevent a value being returned sometimes function LAG...

    Note: I'll have to deal with a database 8i on this one...

    Here is a simplified table & data:
    CREATE TABLE     steps
(     item_id          CHAR(25)
,     ord_nbr          CHAR(10)
,     sub_nbr          CHAR(3)
,     step_nbr     CHAR(4)
,     start_date     DATE
,     finish_date     DATE
);

INSERT INTO steps
VALUES ('A','0000000001','001','0010',TO_DATE('01/01/2011','mm/dd/yyyy'),TO_DATE('01/02/2011','mm/dd/yyyy'));
INSERT INTO steps
VALUES ('A','0000000001','001','0020',TO_DATE('01/03/2011','mm/dd/yyyy'),TO_DATE('01/03/2011','mm/dd/yyyy'));
INSERT INTO steps
VALUES ('A','0000000001','001','0030',TO_DATE('01/05/2011','mm/dd/yyyy'),TO_DATE('01/06/2011','mm/dd/yyyy'));
INSERT INTO steps
VALUES ('A','0000000001','002','0010',TO_DATE('01/01/2011','mm/dd/yyyy'),TO_DATE('01/02/2011','mm/dd/yyyy'));
INSERT INTO steps
VALUES ('A','0000000001','002','0020',TO_DATE('01/04/2011','mm/dd/yyyy'),TO_DATE('01/04/2011','mm/dd/yyyy'));
INSERT INTO steps
VALUES ('A','0000000001','002','0030',TO_DATE('01/06/2011','mm/dd/yyyy'),TO_DATE('01/07/2011','mm/dd/yyyy'));
INSERT INTO steps
VALUES ('B','0000000002','001','0005',TO_DATE('01/10/2011','mm/dd/yyyy'),TO_DATE('01/12/2011','mm/dd/yyyy'));
INSERT INTO steps
VALUES ('B','0000000002','001','0025',TO_DATE('01/18/2011','mm/dd/yyyy'),TO_DATE('01/19/2011','mm/dd/yyyy'));
    The actual data are the result of a query and has about 200K lines, but this should be enough to solve my problem... It contains a list of commands and the steps that each order has to pass before a product is finished. There is also the date of beginning and end for each step.

    My ultimate goal is to be able to calculate the product time (days) expected between steps. My first step to get what was:
    SELECT     s.*
,     LAG(s.finish_date)     OVER     (
                         ORDER BY     s.item_id
                         ,          s.ord_nbr
                         ,          s.sub_nbr
                         ,          s.step_nbr
                         )                         AS last_step_finished
FROM     steps s
    This allows me to bring to the date of completion of the previous stage, while in the next level of the query, I get the difference between the previous step finishing and the present stage from give me the waiting time.

    The only problem is, I need for the query return null for the LAG function if the line is the first step in an agenda - a single order is given by ord_nbr + sub_nbr - because I won't be calculate the time between the end of another way and the beginning of another. I will assume that there are no queues in any order. So, I guess I could return NULL at this level, or at a higher level, simply ignore the column last_step_finished if I determine that the rank be the beginning of an order, but whatever it is, I have some difficulty to reconstruct the code to determine if a line is the first step in an order.

    My first thought is that the solution will look something like:
    SELECT     s.*
,     CASE
          WHEN     --we're on the first step of an order
          THEN     NULL
          ELSE     LAG(s.finish_date)     OVER     (
                                   ORDER BY     s.item_id
                                   ,          s.ord_nbr
                                   ,          s.sub_nbr
                                   ,          s.step_nbr
                                   )                    
     END     AS last_step_finished
FROM     steps s

    Hello

    Thanks for being so clear on your version and display the sample data; This is really useful!
    Don't forget to post the exact results you want from these data.

    You don't want to ORDER BY all these columns; you want to PARTITION BY som of them, instead, like this:

    SELECT     s.*
,     LAG (s.finish_date)     OVER     ( PARTITION BY  s.item_id
                           ,          s.ord_nbr
                           ,          s.sub_nbr
                           ORDER BY     s.step_nbr
                         )     AS last_step_finished
FROM     steps s

    The output of this product that is:

    ITEM_ID ORD_NBR    SUB STEP START_DAT FINISH_DA LAST_STEP
------- ---------- --- ---- --------- --------- ---------
A       0000000001 001 0010 01-JAN-11 02-JAN-11
A       0000000001 001 0020 03-JAN-11 03-JAN-11 02-JAN-11
A       0000000001 001 0030 05-JAN-11 06-JAN-11 03-JAN-11
A       0000000001 002 0010 01-JAN-11 02-JAN-11
A       0000000001 002 0020 04-JAN-11 04-JAN-11 02-JAN-11
A       0000000001 002 0030 06-JAN-11 07-JAN-11 04-JAN-11
B       0000000002 001 0005 10-JAN-11 12-JAN-11
B       0000000002 001 0025 18-JAN-11 19-JAN-11 12-JAN-11

    If this isn't what you want, move some of the columns of the PARTITION BY clause from the ORDER BY clause.
    PARTITION OF x, y, z is the analytical consideration of the aggregation GROUP BY x, y, z. This means that each separate combination of x, y and z will be considered to be a clean world, which do not interact with other values of x, y or z. On the first line of each partition, LAG return the NULL value, exactly as you wish.
    Sorry, I don't have Oracle 8 to test, but I'm sure that LAG has not changed since Oracle 8.1.

    Moreover, the CHAR data type is nothing trouble. Use rather VARCHAR2.

  • by using the LAG() and MIN() functions

    So, I try to get the smallest number in a column of an alias I use LAG() in. I know, which makes no sense, so here's what im trying to do with a SQL statement:

    Select MIN (NUM1 - lag (num1, 1, 0)) over (ORDER BY NUM1) AS J_LAG ORDER BY HEY HEY

    Basically I want the smallest number of the column that I run the lag function on. When I try to it says, ' ORA-30484: lack of specification of the window for this function. I tried to put in brackets the minutes like this: MIN (NUM1 - lag (num1, 1, 0)) and as follows: MIN (NUM1 - lag (num1, 1, 0) over (ORDER BY NUM1)) and none of them does not seem to work.

    Any ideas?

    -Joshua 
    like this ??

Select Min(hey) From (Select num1 - lag(num1,1,0) over(ORDER BY NUM1) As hey From J_LAG)

  • Analytic function

    Dear masters,

    Oracle version: 11.2.0.2.0

    I have a table called price_detail where the data is there in the format below

    Sell_date(mm/dd/yyyy)

    Dealer_Price

    Discount

    Net price

    Net_Price_before_discount

    01/01/2013

    100

    10

    02/01/2013

    120

    20

    03/01/2013

    200

    10

    04/01/2013

    100

    20

    05/01/2013

    120

    10

    06/01/2013

    210

    30

    07/01/2013

    140

    40

    08/01/2013

    150

    10

    09/01/2013

    200

    20

    10/01/2013

    100

    10

    11/01/2013

    120

    30

    12/01/2013

    140

    20

    In my program, I am passing a date range (for this example say entry date between 04/01/2013 09/01/2013), I need to calculate the NET_PRICE and the NET_PRICE_BEFORE_DISCOUNT for a given date range

    Sell_date(mm/dd/yyyy)

    Dealer_Price

    Discount

    Net price

    Net_Price_before_discount

    01/01/2013

    100

    10

    02/01/2013

    120

    20

    03/01/2013

    200

    10

    04/01/2013

    100

    20

    70

    100

    05/01/2013

    120

    10

    80

    190

    06/01/2013

    210

    30

    200

    70

    07/01/2013

    140

    40

    120

    80

    08/01/2013

    150

    10

    140

    200

    09/01/2013

    200

    20

    170

    120

    10/01/2013

    100

    10

    11/01/2013

    120

    30

    12/01/2013

    140

    20

    Net = Dealer_price - Discount price (of the 3rd day after the date of the day, so if we are calculating for 04/01/2013 then delivery must come from 06/01/2013)

    Net_ Price_ before_ discount = Dealer_price (of the 3rd day before the current day, so if we are calculating for 04/01/2013 then the Dealer_price must come from 02/01/2013) - Discount

    I can achieve this LEAD and functions LAG problem I'm facing is that I need to calculate it only for a date range and some values (as shown above) fall not going so here it is coming as 0.

    1. for example when I am Net_price calculation for the 09/01/2013, the discount is from 11/01/2013, but as in my case clause I have a widow who allows me to get data only between 01/04/2013 and on 09/01/2013, the discount is coming up as 0.

    It's really complicated for me to achieve.

    Please suggest.

    AB

    Every time that you give examples of data give as CREATE TABLE and INSERT QUERIES or the WITH clause. Its easy when its this way rather than the data in a structure of table posted by you.

    You can use LEAD and LAG function like that. You must specify the OFFSET value to meet your need.

    with t

    as

    (

    Select to_date (' 01 / 01/2013 ', ' dd/mm/yyyy') sell_date, dealer_price 100, 10 double discount

    Union of all the

    Select to_date (' 02 / 01/2013 ', ' dd/mm/yyyy') sell_date, dealer_price 120, 20 double reduction

    Union of all the

    Select to_date (' 03 / 01/2013 ', ' dd/mm/yyyy') sell_date, dealer_price 200, 10 double discount

    Union of all the

    Select to_date (' 04 / 01/2013 ', ' dd/mm/yyyy') sell_date, dealer_price 100, 20 double reduction

    Union of all the

    Select to_date (' 05 / 01/2013 ', ' dd/mm/yyyy') sell_date, dealer_price 120, 10 double discount

    Union of all the

    Select to_date (' 06 / 01/2013 ', ' dd/mm/yyyy') sell_date, 210 dealer_price, 30 double discount

    Union of all the

    Select to_date (' 07 / 01/2013 ', ' dd/mm/yyyy') sell_date, dealer_price 140, 40 double reduction

    Union of all the

    Select to_date (' 08 / 01/2013 ', ' dd/mm/yyyy') sell_date, dealer_price 150, 10 double discount

    Union of all the

    Select to_date (' 09 / 01/2013 ', ' dd/mm/yyyy') sell_date, dealer_price 200, 20 double reduction

    Union of all the

    Select to_date (' 10 / 01/2013 ', ' dd/mm/yyyy') sell_date, dealer_price 100, 10 double discount

    Union of all the

    Select to_date (' 11 / 01/2013 ', ' dd/mm/yyyy') sell_date, dealer_price 120, 30 double discount

    Union of all the

    Select to_date (' 12 / 01/2013 ', ' dd/mm/yyyy') sell_date, dealer_price 140, 20 double reduction

    )

    Select sell_date

    dealer_price

    discount

    , case when sell_date between to_date (' ' 04 / 01/2013 ', ' dd/mm/yyyy ') and

    TO_DATE (' 09 / 01/2013 ', ' dd/mm/yyyy "")

    then dealer_price - lead (2 discount) (sell_date order)

    Another null

    end net_price

    , case when sell_date between to_date (' ' 04 / 01/2013 ', ' dd/mm/yyyy ') and

    TO_DATE (' 09 / 01/2013 ', ' dd/mm/yyyy "")

    then lag (dealer_price, 2) during delivery (order of sell_date).

    Another null

    end net_price_before_discount

    t;

    Output...

    SELL_DATE DEALER_PRICE NET_PRICE NET_PRICE_BEFORE_DISCOUNT DISCOUNT

    --------- ------------ -------- --------- -------------------------

    1ST JANUARY 13 100 10

    2 JANUARY 13 120 20

    3 JANUARY 13 200 10

    4 JANUARY 13 100 20 70 100

    5 JANUARY 13 120 10 80 190

    6 JANUARY 13 210 30 200 70

    7 JANUARY 13 140 40 120 80

    8 JANUARY 13 150 10 140 200

    9 JANUARY 13 200 20 170 120

    10 JANUARY 13 100 10

    11 JANUARY 13 120 30

    JANUARY 12, 13 140 20

    12 selected lines

  • help with LAG () AND dates

    I have at this moment the following query
      SELECT "DATE",
          "CELL_SITE",
          "LASTV_ATTCNT",
          "LASTV_ATTCNT2",
          "LASTV_BLKCNT",
          "LASTV_DRPCNT",
          "V_ATT_CNT",
          "V_CUST_BLK_CNT",
          "V_DRP_CALL_CNT",
          "LASTD_ATTCNT",
          "LASTD_BLKCNT",
          "LASTD_DRPCNT",
          "D_ATT_CNT",
          "D_CUST_BLK_CNT",
          "D_DRP_CALL_CNT"          
     FROM (  SELECT DATE,
                   CELL_SITE,
                    LAG (SUM (V_ATT_CNT), 24)
                       OVER (PARTITION BY BSM_NM ORDER BY DATE)
                       AS "LASTV_ATTCNT",
                    LAG (SUM (V_CUST_BLK_CNT), 24)
                       OVER (PARTITION BY BSM_NM ORDER BY DATE)
                       AS "LASTV_BLKCNT",
                    LAG (SUM (V_DRP_CALL_CNT), 24)
                       OVER (PARTITION BY BSM_NM ORDER BY DATE)
                       AS "LASTV_DRPCNT",
                    LAG (SUM (D_ATT_CNT), 24)
                       OVER (PARTITION BY BSM_NM ORDER BY DATE)
                       AS "LASTD_ATTCNT",
                    LAG (SUM (D_CUST_BLK_CNT), 24)
                       OVER (PARTITION BY BSM_NM ORDER BY DATE)
                       AS "LASTD_BLKCNT",
                    LAG (SUM (D_DRP_CALL_CNT), 24)
                       OVER (PARTITION BY BSM_NM ORDER BY DATE)
                       AS "LASTD_DRPCNT",
                    SUM (V_ATT_CNT) AS "V_ATT_CNT",
                    SUM (V_CUST_BLK_CNT) AS "V_CUST_BLK_CNT",
                    SUM (V_DRP_CALL_CNT) AS "V_DRP_CALL_CNT",
                    SUM (D_ATT_CNT) AS "D_ATT_CNT",
                    SUM (D_CUST_BLK_CNT) AS "D_CUST_BLK_CNT",
                    SUM (D_DRP_CALL_CNT) AS "D_DRP_CALL_CNT"
               FROM DMSN.DS3R_FH_1XRTT_FA_LVL_KPI
              WHERE DATE >
                         (SELECT MAX (DATE) FROM DMSN.DS3R_FH_1XRTT_FA_LVL_KPI)
                       - 2
           GROUP BY DATE, CELL_SITE)
    WHERE DATE >=
             (SELECT MAX (DATE) - NUMTODSINTERVAL (12, 'HOUR')
                FROM DMSN.DS3R_FH_1XRTT_FA_LVL_KPI)
    What I've noticed, is that the LAG function is kind of what I want but not exactly. Let's say that I have given for the hours of 12: 00, 01:00, 02:00, 03:00 04:00, 05:00, 06:00, 07:00, 08:00, 09:00, 10:00, 11:00, 12:00 and the time is 12:00, the LAG function goes back to 12 h, which is what he should do.

    but lets say that I have given for 12 h, 01:00, 02:00, 03:00 04:00, 05:00, 07:00, 08:00, 09:00, 10:00, 11:00, 12:00, but it me MISSING at 06:00 (no data from 06:00). The lag function now back at 23:00 instead of 12 AM because the time 06:00 is missing.

    If I have data for 12: 00 01:00, 02:00, 03:00 04:00, 07:00, 08:00, 09:00, 10:00, 11:00, 12:00 and am MISSING 05:00 and 06:00, the data will return to 22:00 then and not only from 12: 00.

    Can I prevent this and always just come back in 12 hours?

    Published by: k1ng87 on April 25, 2013 13:27

    Hello

    k1ng87 wrote:

    ...
LAG (SUM (V_ATT_CNT), 24)
OVER (PARTITION BY BSM_NM ORDER BY DATE)
AS "LASTV_ATTCNT",
LAG (SUM (V_CUST_BLK_CNT), 24)
OVER (PARTITION BY BSM_NM ORDER BY DATE)
AS "LASTV_BLKCNT",
LAG (SUM (V_DRP_CALL_CNT), 24)
OVER (PARTITION BY BSM_NM ORDER BY DATE)
AS "LASTV_DRPCNT",
...

    Since you want to so many columns of the same previous row, it would be probably easier and more effective to make a self-join, rather than so many functions of SHIFT.

    What I've noticed, is that the LAG function is kind of what I want but not exactly. Let's say that I have given for the hours of 12: 00, 01:00, 02:00, 03:00 04:00, 05:00, 06:00, 07:00, 08:00, 09:00, 10:00, 11:00, 12:00 and the time is 12:00, the LAG function goes back to 12 h, which is what he should do.

    You always pass 24 as 2nd argument offset; What does "Return the value of the line 24 of the precding." If there is a line for each hour, the current line is from 12:00, then wouldn't 24 ranks back would be from 12:00 yesterday, not 12 AM? Maybe you wanted to spend 12, not 24, as the second argument of LAG. I'll assume you really want the 24 line rom now on, but the technique is the same that that number either 24 or 12.

    but lets say that I have given for 12 h, 01:00, 02:00, 03:00 04:00, 05:00, 07:00, 08:00, 09:00, 10:00, 11:00, 12:00, but it me MISSING at 06:00 (no data from 06:00). The lag function now back at 23:00 instead of 12 AM because the time 06:00 is missing.

    If I have data for 12: 00 01:00, 02:00, 03:00 04:00, 07:00, 08:00, 09:00, 10:00, 11:00, 12:00 and am MISSING 05:00 and 06:00, the data will return to 22:00 then and not only from 12: 00.

    Can I prevent this and always just come back in 12 hours?

    A self-join would handle this automatically: it will look for the corresponding line, without regard for other lines may be present or absent.

    If you really had to do with the analytical functions, you can use MIN or FIRST_VALUE with a window of the range:

    RANGE BETWEEN  1 + (.5 / 24)  PRECEDING
      AND      1 - (.5 / 24)  PRECEDING

    DATE when you ORDER BY a column (do not call ' DATE'), the units are 1 day, so 1 + (.5/24) days ago is 24.5 hours; in other words, the window includes only DATES between 24.5 and 23.5 hours.

    If, for some reason, you really want to call LAG, so you could do an outer join to ensure that a row had been present for each of the previous 24 hours.

    I hope that answers your question.
    If not, post a small example data (CREATE TABLE and only relevant columns, INSERT statements) and also publish outcomes from these data.
    Explain, using specific examples, how you get these results from these data.
    Always say what version of Oracle you are using (for example, 11.2.0.2.0).
    See the FAQ forum {message identifier: = 9360002}

  • With the help of lag and grade in the same query

    Hello

    I'm trying to find out the time difference between the memberships peoples, as well as the order in which these memberships are in. So far, I have added in a release status of work on the order, in that the accessions were created, but now want to look at the difference between the dates returned. The SQL, I used is:

    SELECT owner_party_id,
    mem_number,
    support_id,
    mem_start_date,
    Rank() over (PARTITION BY owner_party_id ORDER BY mem_start_date ASC) MEMBERSHIP_SEQUENCE
    OF membership_all
    WHERE version_type = "NEWS".
    AND owner_party_id IN ('65051', '65051', '65348', '65348', '65607' 65607', 65607')

    to get:

    "OWNER_PARTY_ID" | "' MEM_NUMBER ' | "' SUPPORT_ID ' | "' MEM_START_DATE ' | "" MEMBERSHIP_SEQUENCE ".
    65051 | 318874751 | 8014747 | 1ST MARCH 10 | 1
    65051 | 412311060 | 21502883 | 15 AUGUST 12 | 2
    65348 | 308672459 | 3526913 | MAY 1, 10. 1
    65348 | 409951130 | 20950524 | 18 JUNE 12 | 2
    65607 | 315830192 | 7510133 | MAY 17, 10 | 1
    65607 | 406448110 | 20024246 | MARCH 16, 12. 2
    65607 | 409738130 | 20903556 | JUNE 14, 12. 3

    Now I want to calculate the difference between the starting dates of each of the groups owner_party_id, in order to get something like this:

    OWNER_PARTY_ID | MEM_NUMBER | SUPPORT_ID | MEM_START_DATE | MEMBERSHIP_SEQUENCE | Diff
    65051 | 318874751 | 8014747 | 1st March 10 | 1.
    65051 | 412311060 | 21502883 | 15 August 12 | 2. 898
    65348 | 308672459 | 3526913 | May 1, 10. 1
    65348 | 409951130 | 20950524 | 18 June 12 | 2. 779
    65607 | 315830192 | 7510133 | May 17, 10 | 1
    65607 | 406448110 | 20024246 | March 16, 12. 2. 669
    65607 | 409738130 | 20903556 | June 14, 12. 3. 90

    I think I need to use the Lag function, but I'm not sure if it can be linkited looking at the data within a group id of owner of party, that it would be foolish to calculate the difference in dates for the two different owner group ID.

    Any advice much appreciated.

    Thank you

    Published by: 992871 on March 9, 2013 23:34

    Maybe NOT TESTED!

    SELECT owner_party_id,
       mem_number,
       support_id,
       mem_start_date,
       RANK() OVER (PARTITION BY owner_party_id ORDER BY mem_start_date) MEMBERSHIP_SEQUENCE,
       mem_start_date - LAG(mem_start_date,1) OVER (PARTITION BY owner_party_id ORDER BY mem_start_date) diff
  FROM membership_all
 WHERE version_type = 'CUR'
   AND owner_party_id IN ('65051','65051','65348','65348','65607','65607','65607')

    Concerning

    Etbin

Maybe you are looking for