Time series OBIEE 11 g: regression on there / function to this day?

Hello

We test the migration to OBIEE 11 G.
We are currently in version 10.3

We have a problem on AGO / OBIEE TO_DATE functions that returns null results on existing reports.
In version 10.3, these functions operate properly with the real TIME dimension.

Here's how we designed the time Dimension:

TOTAL-> year-> shift-> MONTH->-> DATE_KEY, FULL_DATE WEEK

DATE_KEY is a number and referenced on our fact table as a foreign key.

CREATE TABLE TD_DATE
+(+
DATE_KEY NUMBER (11) NOT NULL,
DATE OF FULL_DATE
THE WEEK NUMBER (2),
NUMBER OF THE MONTH (2),
NUMBER OF QUARTER (1)
NUMBER OF THE YEAR (4).
TOTAL_DATE VARCHAR2 (11 BYTES)
+)+

Database version: Oracle 11 g

All items in the dimension are chronological touch (except full date) to optimize performance.

Have you encountered similar problems in inplementation of AGO / TO_DATE function on the 11 g version?

Thank you

Roman

A month have always the same number of weeks and begin with the first day of a week?
If it is not the case, your dimension is not good when you're doing a report on the grain of the week for a year, for example.

Here's a checklist:
http://gerardnico.com/wiki/dat/OBIEE/ago_todate_configuration#timecalender_dimension_design_verification

Success
Nico

Tags: Business Intelligence

Similar Questions

  • There is function to this day do not using aggregates

    I use the Group table and a detailed table
    I have the time hierarchy, which is the lowest level to date. I have for sale
    created for a year (CDA) comparison with function AGO. He
    works well.

    BUT, I want to accelerate, so I used aggregates. I created new
    table time grouped by MONTHS and table corresponding agg_SALES. I have
    established connections and the mapped fields. It works OK for
    normal queries, this means that my questions to use month level
    aggregates. BUT when I add YAGO measure or measures of the CDA in the query, it will
    then the table sales whose grain is date.

    How can I force BI Server to use the AGO with aggregates?

    I use OBIEE 10.1.3.4

    Thank you

    Ishaq

    (Question originally shot of ITtoolbox and published by Gorazd)

    Hello

    By defintion TODATE and function there is need the lowest granularaty of the function. Alternativily, you can write you own function on the database using your aggregates. and use the EVALUATE function in OBIEE to push to the interface.

    Concerning

    John
    http://obiee101.blogspot.com/

  • There are to this day/function

    How to use AGO/TODATE function in response not RPD?
    I do not know how to insert data, such as time_level [there is (expr, time_level, offset)]

    Hey Kobe,

    If your

    1. subject domain name is: SAN

    2 dimension folder_name is: DFN

    3 hierarchy name is: HN

    4. 1st level of the name of the hierarchy is: THN

    then your function Ago syntax is something like

    There are ('measures'. "" Measure1 ', 'SAN '. "" DFN ". "" HN ". ("" THN ", 1).

    I hope this helps.

    Thank you
    Diakité

  • Doubts about time series measures

    Hello

    In my case, I have two calendars

    (1) Calendar Fiscal Enterprise (gold) (August to July)

    2) of the Gregorian calendar (Jan-Dec).

    So now my question is if we apply the function to this day using two calendars, how it will work.

    For ex: if I create a function of year to date using the Gregorian calendar, then it will pick up the results of Jan. 1 so far until. And even if I used the factory calendar then it will pick up the results of August 1 so far until.

    Is my above correct understanding. Correct me if am wrong anywhere.

    And now I have a requirement, my client wants to change the dates in the calendar, I mean for the Fiscal Calendar, they won't leave (August to July), but (Dec-Dec) is this possible, if so please specify how and where...


    Help is greatly appreciated.

    Raj,
    For your condition, I've written a procedure in which the date of 1 January 2010 ' he converted to the exercise of December 26, 2009 "and the same is updated within your calendar_table in the fis_dat column based on which you can easily find your fiscal_year, month and so on." The same can be much simplified and I'll update if find the time... Check and to go me back.

    Create or replace procedure fin_cal
    as
    Start
    DECLARE
    date of temp_date;
    date of fisc_date;
    date of temp_fisc_date;
    date of previous_date;
    int pre_year;
    int year_date;
    cursor c1 is select trunc (date_time) of dates1 where time_key > = 20100101 and time_key<= 20120101="" order="" by="" trunc(date_time);="" calculated="" from="" 1st="" jan="" 2010="">
    Start
    Open c1;
    loop
    Fetch c1 into temp_date;
    When exit c1% notfound;
    year_date:=TO_NUMBER(to_char(to_date(temp_date,'dd-mm-yy'),'yyyy'));
    pre_year: = year_date-1;

    If to_char (temp_date, 'dd') = '01' and to_char (temp_date, 'mm') = '01' and then
    fisc_date: = to_date(pre_year||) e 12-24', 'YYYY-MM-DD'); the initial date is 24-12-title...
    dbms_output.put_line (fisc_date);
    Update dates1 set fis_dat = fisc_date where trunc (date_time) = trunc (temp_date); fis_dat is the column in the table
    on the other

    Select fis_dat in the previous_date of dates1 where trunc (date_time) = trunc (temp_date)-1;
    temp_fisc_date: = to_date (to_char(previous_date,'YYYY-mm-DD'), 'YYYY-mm-DD') + 1;
    Update dates1 set fis_dat = temp_fisc_date where trunc (date_time) = trunc (temp_date);
    temp_fisc_date: = ";
    end if;
    end loop;
    end;
    end;

    Start
    fin_cal();
    end;

    Kind regards
    Anitha.B

  • Time series by using Oracle Data Mining? Is there any equivalent of ARIMA in the MDGS?

    Hello

    I followed the extraction of data from Oracle Press book. I applied regression models to predict continuous variables.

    My use case is to predict the market price of stocks based on historical data.

    I see there R packages to do. Some popular methods that are widely used are ARIMA (Autoregressive integrated mobile average).

    Is there such an offer out-of-the-box in the MDGS?

    Pointers / whitepapers, or do we use component regression?

    Thank you

    Chrystelle

    Hey Lulu,

    MDG does not have an implementation of time series which is precisely part of the SQL in the algorithms of db.

    But AAO (Oracle Advanced Analtyics, ODM is a part of) has an implementation of time-series Oracle which may be useful for you:

    See ore.esm model, documented with examples here: https://docs.oracle.com/cd/E57012_01/doc.141/e56973/procs.htm#BEIEGCDG

    We also have the preparation of data for time series capabilities, listed here: https://docs.oracle.com/cd/E57012_01/doc.141/e56973/procs.htm#BEIJHHCG

    May also be interested to Marcos Campos blog on time series using the current characteristics of ODM/DB: Oracle Data Mining and Analytics: time series Revisited

    Hope it will be useful.

    THX, mark

  • Time series: finding max and min for each month

    Hello world. I'm working on a datasheet that has collected the data points for all day for the past 15 years.

    A column stores the date and column B contains a numeric value.

    I want to do the following:

    Create a new table (call it table #2) that:

    (1) is now a time series on a monthly basis (column A contains "year month"). This is why the table contains a header and 15 * 12 = 60 rows of data

    (2) B Col contains the Max value of this month.

    (3) C Col contains the value this month min

    Table #2 should be calculated automatically because I need to find these values for several series. Any ideas?

    Here is a way that I came up with

    In the table above, your original data is in columns B and C.

    I've added three additional columns (A, D and E):

    D2 = Year (B2) & NUMTOBASE (MONTH (B2), 10, 2)

    It's shorthand dethrone select D2, and type (or copy and paste it here) the formula:

    = Year (B2) & NUMTOBASE (MONTH (B2), 10, 2)

    E2 = IF (D1≠D2, 0, E1 + 1).

    Select copy of D2 to E2,

    Select cells D2 at the end of column E, block

    A2 = D2 & NUMTOBASE (E2, 10, 2)

    Select A2, copy

    Select A2 at the end of the A2 column, paste

    now create a second table summary as shown:

    For 'Summary Table 2016:

    the first four lines are the lines of header

    There are 35 total lines

    Enter the year in cell B1

    A5 = −4 LINE)

    B5 = SIERREUR (VLOOKUP ($B$ 1 & NUMTOBASE(B$4,10,2) & NUMTOBASE($A5,10,2), data: $A:$ C, 3, 0), "")

    Select cell B5, copy

    Select cells B5 at the end of row 5, dough

    Select A5 thru M5, copy

    Select cell A5 at the end of the M column, paste

    B2 = MIN (B)

    B3 = MAX (B)

    Select the cells B2 to B2, copy

    Select the cells B2 through M3, dough

    now duplicate the table and change the year to 2015 for the next year.

    You can duplicate this table as necessary to summarize a year

  • Best approach to model Time series Calc

    Hello comrades,.

    Please let us know the best way to use the calculation of time series in the reports (IE last YEAR, BAT etc.).

    Since many developers odious at this stage

    to share your ideas and suggestions how to model side RPD and reports, while it would be easy to avoid any mistakes or false calculations.

    Thanks in advance...

    Hello

    The easiest is to simply use the provided time series functions out of the box: to use it all you need to do is set up correctly your time dimensions (levels and keys) and the function will do the job.

    Another option may be to model the time series you need as physical join and amounts by a natural aggregations of the rows returned. For this one, you have 2 options: create the values on the side on the side of time dimension or fact. Depending on how you will you need to add additional columns in your MDB to represent the values of time series.

    It really depend on several things to decide which option to use. Start with the output of the function box and there are many chances it will be just enough for most needs.

  • Functions of time series

    I added a time series measured in SPR

    Year ago sales = there is ("Business Analytics" "." ") Sales facts ".» "The sale $", "Business Analytics". «Time hierarchy ".» ("" Exercise ", 1).

    If I shoot "Sales there of the year" in the answers, it takes forever to run - (5 - 6 minutes) - the result is correct.

    If I simply pull - "Business Analytics". "Facts sale '." "" Selling the $'s ' running in 5 seconds.

    Is their anyway I can optimize the timeseries functions?

    Any response would be greatly appreciated!

    We are the Oracle Business Intelligence 11.1.1.6.5 version.

    Published by: ann 2012 March 14, 2013 14:13

    That's why I often make an alias of the fact table and join the in the time dimension with year - 1. For me the SQL out readable and performs generally better than the version that comes with AGO.

    Scott

  • Dynamic time series in a form

    It seems to allow me to add it. Click on the Member, A-T-J of medium-sized options. He then Let's save me it.

    But then not let me open the form (errors).

    Anyone know if this is a bug or if there is something missing.
    DTS works in reports
    I use version 11.1.2.1

    It may depend on the form design
    for example 10226960 - cannot open a data form that includes members of the time series dynamics (DTS) and has the option to remove the missing blocks selected.

    See you soon

    John
    http://John-Goodwin.blogspot.com/

  • Time series graphics display data inappropriate for a continuous analysis

    Hi all
    I've marked in months as chronological key in my MDB layer, but I still cannot display data correctly in my chart of time series, because it shows data for analyses continuously when you create the chart. Can someone help me with the same.


    Thank you

    What type of data is your key? The chronological key is required for time series formulas (there etc.).

    The time series graph requires a date or datetime data type work - maybe a new column with the first of the month / help? Kind regards

    Robret

  • Question of function of time series

    Hi all

    I use the function of time series for year there are recipes, in fact table, the date is 2007 to 2010, but related,.

    It has a row of 2011, 2011 income is null and income year is 2010, report as follows:
    -----
    income for the year year_ago_Revenue
    2007 __50___
    2008 __80___ 50
    __100___80 2009
    2010 __120___100
    ___120 2011

    I want to delete the line of the year 2011, how can I do this?

    Since you only ask how to get rid of the line of 2011, here are two ways:

    (1) filter on column Time.Year <> 2011

    or better

    (2) filter on column Time.Year <> YEAR (CURRENT_DATE) + 1

  • Conversions of time in OBIEE

    Hi all

    Face problems with conversions from hour to OBIEE. We use MySql database that stores the time in seconds.

    I used the following function to change in a meaningful format and created a new field in the database.

    Select from_unixtime (dateline) in the s swtickets;

    My question is, is there any function available in OBIEE administrator to change the format from seconds to the Date format.


    Thanks in advance,

    Imtiaz.

    Hi Imtiaz,

    Perhaps you could call the from_unixtime through EVALUATE function.

    Good luck

    Daan Bakboord

  • I got a message that there is an hour: Trojan.Win32.Generic detected in the time machine.  How can I remove it? Is this true?

    I received a message that there is an hour: Trojan.Win32.Generic detected in the time machine.  How can I remove it? Is this true?

    I have the version of yosemite 10.10.5 OSX macbook air.

    Whence this message?

  • Dynamic members time series (In a BSO Application) can be used in other formulas of Member, rules of business or Calc Scripts?

    I am building an application of OSI to refer to values for a YEAR in several calculations. Before you enable the DTS, I wanted to make sure that they can be referenced in the formulas (used in IF conditions to check the values against a value of Member A-T-D)

    Dynamics of time series can be used only to display values DTS EN reports or in smart view retrieves. You cannot use them in the form of Member or Calc. I think the best way would be to add a version with a formula that adds every month until this month for CDA or a different scenario.

  • I bought the first elements 13 and 13 Photoshop elements according to a trail. I received 1 set of serial numbers to activate Photoshop and 2 sets for first. However, after you type in the first series of numbers for the first, there was no place to type

    I bought the first elements 13 and 13 Photoshop elements according to a trail. I received 1 set of serial numbers to activate Photoshop and 2 sets for first. However, after you type in the first series of numbers for the first, there is no place to type in the second set. How can I check if the first has been correctly activated?

    You get usually just a serial number by program... a list of 24 numbers, 6 sets of 4.

    Click the icon of the program and see if it boots without asking for a serial number.

    If it is not running now, please contact support:

    To contact support, please click the link below, click on the still need help? option in the blue box below, then choose the cat. Make sure you are signed in with your Adobe ID, have cookies enabled and have deleted your cookie cache.  If you have any questions, try another browser.

    Serial number and activation support

    Guinot

Maybe you are looking for