Time Dimension- Wk and Month

Hi,Our user would like to see sales performance by time which rolls up to different hierarchy paths, with measures which use dynamic time series, such as MTD, YTD.TIME1) Year --> Qtr --> Month2) Year --> Week of yearMy question is since I have two different hierachy paths, should I put these time dimensions in one cube. Thus, the cube would contains the following dimensions:TIMEinMonth, TIMEinWeek, Measures, Prod, Geogaphy Or should I build two cubes: one has time dimension as in months, and another has time dimension as in week.It looks like the first alternative has lower space requirements overall, but I can't tag both TIMEinMonth and TIMEinWeek as "Time" dimension in OLAP outline. Can I still have YTD for both time dimensions?What should be the best design?Thanks.Sam

Why don't you try with the two different time hierachies under the same dimension?TIME Dimension: Time --> Year_Month --> Qtr --> Month | --> Year_Week --> Week of year Do not aggregate the Year_Week to the Year_Month and you will have the Year_Month data at Time level! In this way your cube would contains only the following dimensions: TIME, Measures, Prod, Geogaphy.I have used this solution several times, but sincerely I haven't tested too much the Dynamic time series functions with this architecture.

Similar Messages

Seeking advice regarding Time Dimension - Correction and Customization

Dear all:
Situation
1. We have a measure called HTD, Half Year to Date, which aggregates 6 months worth of data. However, I just realized that the previous consultants set HTD at LEVEL = "Week". It works fine but I feel that this naming convention is confusing, and would like to rename HTD's LEVEL.
2. We have 4 more additional periods after period 12 for adjustment purpose. What is the good practice to have a year with 16 periods?
Data
Current LEVEL for Half Year to Date is named "Week". I would like to rename this to something like "HalfYr"
200x.01 - 12 are regular periods
200x.13 - 16 are additional periods
Question
1. If I go to SQL's back end and rename my LEVEL, and then update my Time dimension, is this going to pose a problem to my historical data?
2. I would like to include additional 4 more periods to sync with BW, but at the same time not confuse my BPC users in running reports at month of December. In other words, how do I integrate the 16 period time dimension design into the concept of "1 yr = 12 months (periods)"?

Tim:
Sorry for the late reply.
We are running BPC 5.1 SP3 and SQL 2003
Yes, MONTH, QUARTER, YEAR are used for LEVEL property, and the only irrelevance is HALF YEAR which has WEEK in LEVEL property.
I think perhaps a separate data source makes sense... Have you handled any case like this before and how did you manage the "additional" periods?
Thank you! Have a great day!
Sincerely,
Brian

Time dimension wizard and leap year error?

I've created a time dimension (rolap) using the wizard. I started with 2006 for 6 years. I used a fiscal calendar, 445, fiscal year start = jan 1, and month start = monday.
The first 2 months of 2008 are ok. However, OWB ends the third month on March 31, 2008 and it should be March 30, 2008. It starts the next month on April 1, 2008 which is a Tuesday.
I've tested with 10.2 and 11.1.
Anybody seen this or have a solution?
Garrett

Garrett
The start day that you specified as being Monday is in fact the week start day and not the month start day. In most cases these will both always land on the day of week you specify, but there are exceptions.
The Fiscal year is split into Quarters and the start of each Quarter is based on the date of the start of the year, so if you started the year at Jan 6th, then the 2nd quarter will start on April 6th, and the third on July 6th and so on.
In your case, an extra day is inserted in the 5th week of the first quarter so that the quarter can start at the correct date

Setting up a TIME dimension in OBIEE

Hi,
We have an Oracle OLAP Cube built using AWM v10.2.0.3A. We have also created a relational view for this cube (using the OLAP view generator) and are now using this view as a data source in OBIEE v.10.1.3.3.2.
As part of defining the Time dimension in the Business Model Layer of OBIEE repository, we have done all the relevant mappings to the TIME dimension as in the AWM Cube and have marked the dimension as type 'TIME'. Also, for the TIME logical table, we have defined each level (All Years, Year, Month, Quarter, and Day) to have its own chronological key with unique values at all levels. The primary key for our TIME table in the business layer is 'Day'. This level 'Day' is mapped to TIME_DAY_LVLDSC column of the relational view created over the AWM cube. The data type for this column is "VARCHAR". (We basically picked up all the 'Description' columns while creating the view over the cube!)
With this background of the TIME dimension structure, here is the problem scenario:
When we try to create a report with one AGO and one TODATE function defined over the "Day" level (both are defined as logical columns), then the report takes forever to run and never comes back with any data/error even for a handful of rows. We are seeking the forum's help on the following:
1. Is it necessary for the primary key 'Day' of the TIME logical table to be of type DATE? If yes, does this mean that we have to change the column data type in our cube (or select a 'date' type column to be a part of the view and then redo the mapping in the repository) or do we just make a change to the column data type in the physical layer of OBIEE repository?
2. Going by the description above, is there something missing (other than point 1. above) in the setup of the TIME dimension in OBIEE so as to make it usable for Time Series type of calculations/functions?
Any help is much appreciated.
Thanks and Regards,
Piyush

Hi Wildmight,
Thanks a lot for your kind inputs.
1. No, VARCHAR is fine. You should have also defined unique time keys in your time dimension (Quarter Year, Month Year) and check box 'time dimension' in the properties of your time dimension.
-->> The Time dimension that we have defined in the Business layer of the repository is mapped to the respective level columns in the AWM cube and has already got unique members at each level. We are using the "Description (LVLDSC)" columns for each level and have defined the "Description" column to have unique values at each level. A sample data set looks like :
Year:
2007
2008
Quarter:
QUARTER1 2007
QUARTER2 2007
QUARTER1 2008
QUARTER2 2008
Month:
JANUARY 2007
FEBRUARY 2007
JANUARY 2008
FEBRUARY 2008
Day:
Monday, January 1st 2007
Tuesday, January 2nd 2007
So in the above case, do we still need to define a unique key for each level like Quarter Year, Month Quarter, etc.? (Please note that the check box for 'TIME' dimension is checked !)
2. Is your AGO based on TODATE? or your TODATE based on AGO?
-->> I am not sure what this really means - however I can tell you that we have created two new logical columns (in the fact table in Business Layer) : one to show the measure as TODATE for Day level, and the other to show the measure as Day Ago.
So it means that we are trying to create two reports with three columns as below:
- Day Measure on Day Measure Todate (Day)
- Day Measure on Day Measure Day Ago (Day, 1)
I hope I have been able to clarify the problem scenario with this. Does this setup seem correct to you?
Many thanks for your help.
Best Regards,
Piyush

Cannot recreate mapping to populate MOLAP time dimension

I am using OWB10gR2 to create a time dimension having MOLAP as storage type and two hierarchies (year and week).
I used the wizard to create a time dimension having the standard year hierarchy. The wizard does not allow to generate the time dimension with both hierarchies (why?). That's why I had to manually add the week hierarchy with the object editor in a second step. The next step would be to recreate the mapping to populate the time dimension, but I got the following error message:
"There is no sequence attached to the dimension. Please add a sequence in the attribute tab."
I tried the same thing with ROLAP as storage types: for ROLAP I can recreate the mapping.
Does anybody have a solution for that problem?
Many thanks in advance!!

I further explored the Time dimension wizard and for standard if i select year,month and day .... the autogenerated mapping generates a validation error... Anyways i am using oracle11gr2 any ideas why this is happening ??

Sorting Time dimension in EVDRE

Hi all,
This is probably an easy question that for some reason I fail to answer.
I have an EVDRE with Time dimension in the columns.
I have defined my Time dimension as 2011.Jan, 2011.Feb, 2011.Mar etc.
One line below the key I have placed a description (in Hebrew) and I have indicated to hide the column keys.
During the expansion the dimension is not sorted properly. The order (by month number) is:
2011.Feb 2011.Jan 2011.Mar
2011.Apr 2011.Jun 2011.May
2011.Aug 2011.Jul 2011.Sep
2011.Dec 2011.Nov 2011.Oct
It seems that the system sorts it first by querter and then by alphabetic order.
Any ideas how can I control the sorting of columns in EVDRE?
I have also upgraded to 7.5 SP06. I last saved the template in SP05. When I open the template I can see that the dimension is sorted properly and then during the expansion it changes the sorting order. Did any one tackle this in SP06?
Thanks,
Avihay

Hi Avihay,
Actually the dimensions in reports or input schedules appear in the same order as you have maintained the hierarchy. But, yes, you can make some sort of arrangements using control panel memberset option. For eg. you can see parents after or parents before the base level members. There is no other option to change the order of members being displayed in IS or reports.
As far as your issue is concerned, even I was facing exactly the same problem. I checked time dimension membersheet and there found that 'TIMEID' property was not maintained properly. Therefore, would request you to cross-check the TIMEID property in membersheet and if it has been assgned correctly to each of the time dimension members. Goto to Maintain dimension members, change the sheet and reprocess the dimension.
I think this will resolve your issue.
Rgds,
Poonam

Help with Time Dimensions

Hi,
I am trying to create a Time Dimension using the Wizard, but I want to add certain fields to the dimension like Month Ago, Week Ago, Quarter Ago, etc. What are the steps to do this?
Also, is it possible to combine Fiscal and Calendar year into the same dimension?
Thanks

Gurus, anybody anything on this one??
I haave done a small adjustment in the dimension order by which I have eliminated the following error:
Record #40 - Error adding Dynamic calc property to member xxxxxxxxx (3320)
Record #40 - No time dimension defined
But, I have the following new error/warning (cube build successfull in this case) when I made this adjustment:
Record #427 - Error adding Attribute to member PATB4_ADJ-PDA (3320)
The adjustment I made is... I moved TIME dimension up and ACCOUNT dimension below that...so the order I have now is TIME,ACCOUNT,SCENATRIO,XXX,XXXX instead of ACCOUNT,TIME,SCENARIO,XXX,XXXXX ..*would this create any performance issues?*

Composite key in Time dimension

Hi All,
I would like to know Time dimension with Composite key. I have a requirement where I want to store 2 Calendars in Time dimension table. for e.g :
for one calendar Weekstarts from SUN-SAT and for another it is from MON-TUE
DateKey Type WeekStart WeekEnd
20140101 1 Sun Sat
20140101 2 Mon Tue
..................etc
I have a measure group which is related to Time dimension (DateKey and Type used a Composite key). This implementation has no issues for additive measures but there are few issues with semi-additive measures (last non-empty,...).
Will composite Key have any effect on semi-additive measures ?
what if i use surrogate key instead of composite key.
Please suggest if the approach has any issue with Time intelligence. Advise if there is any better approach for the same.
Ram MSBI Developer

Hey.. Thanks!
I am clear about the concept about defining annotation based composite key. Also, I read in the documentation that I'll be needing to define as direct, aggregate or one-to-one. But, I am not able to define and run the same in the project mapping xml of toplink.
It would be great if you can share some sample code for defining the same. For e.g. in my mentioned example, there is TestEntity POJO having 'id' as the attribute which gets populated with the testEntityCode of the TestEntityKey POJO. Please suggest the same for the same:
<opm:primary-key>
<opm:attribute-name>id</opm:attribute-name>
<opm:field table="TEST_ENTITY_B" name="TEST_ENT_CODE" xsi:type="opm:column"/>
</opm:primary-key>
Thanks!

Create time dimension table in repository without data warehouse

Hi,
I want to implement only BI repository solution in my customer (not datawarehousing). Is it possible to transform the data by repository tools, so that the times columns in fact tables are categorized by the "time dimension" table?
To be more explanatory:
The "Sales" table has the "time of sale" column. It contains the timestamp when the sale was performed. I have imported this table in "physical layer" of the repository. Now I want to create a new "time dimension" table, something like:
CREATE TABLE dimension_time (
Day_Key INT NOT NULL PRIMARY KEY,
Day_Timestamp DATETIME NOT NULL,
Day_Name NVARCHAR(32) NOT NULL,
Day_Text NVARCHAR(32) NOT NULL,
INSERT INTO dimension_time VALUES (20110101, {d '2011-01-01'}, '1/1', 'January 1', 'Saturday', 0, 6, 1, 1, 185, 1, 201052, 'W52', 'Week 52', 52, 201101, '01', 'January', 1, 7, 1004, 'Winter', 'Winter', 20111, 'Q1', '1st Quarter', 1, 20103, 'Q3', '3rd Quarter', 3, 20111, 'S1', '1st Semester', 1, 20102, 'S2', '2nd Semester', 2, 2011, '2011', '2011', 2010, '10/11', '2010/2011', 0);
INSERT INTO dimension_time VALUES (20110102, {d '2011-01-02'}, '2/1', 'January 2', 'Sunday', 0, 7, 2, 2, 186, 2, 201052, 'W52', 'Week 52', 52, 201101, '01', 'January', 1, 7, 1004, 'Winter', 'Winter', 20111, 'Q1', '1st Quarter', 1, 20103, 'Q3', '3rd Quarter', 3, 20111, 'S1', '1st Semester', 1, 20102, 'S2', '2nd Semester', 2, 2011, '2011', '2011', 2010, '10/11', '2010/2011', 0);
and after to add a new column in "sales" fact table for "time dimension ID" and through the repository populate this column based on the "time of sale" column and the corresponding "time dimension ID".
I know that the ETL process might perform it, but I do not want to go for Data Warehousing (it is not real - time, needs more resources, etc).
Is it possible to perform such action only on repository?
Thank you.

Hi,
I can do it, but this would be usefull only to create "time dimension" table. But also the "sales" fact table needs to be altered (thus, the "time" column will not contain the value of the time, but the ID of the corresponding time in the "time dimension" table).
I know that on DW this procedure is done automatically by the ETL process.
My question is that does the repository has any tools similar to this?
Thank you.

Loading of time dimension

Hi,
I started thinking about what the best way to load the time dimension is. Until now I have loaded data for 10-15 years in advance into the dimension to be done with it. One down side to this is that when you create or view reports, you might see the year 2015 but for obvious reasons there won't be any data loaded for this year. Another thing I started thinking of the other day is how this may affect the performance and the loadtime of the cube.
If all dimensions is set to not be sparse and you have loaded 10 years that's not yet in use, the size of the cube, from what I understand, will be unnecessarily large. This will probably also have an affect on the loadtime.
If you set all the dimensions except for the time dimension as sparse, as AWM sets it as default, how will this affect the size of the cube?
Anyone have any thoughts around this?
Regards Ragnar

Hi,
I need to load time dimensions to have data till 2015. Moreover we have just regular time dimensions tables and no fiscal tables.
We have updated the fiscal data file to include data till 2015 and update DAc parameter Analysis end date to 31/12/2015. Data is loaded in these tables till 2015 but the ETL is failing for analyze W_ETL_RUN_S task.(param-g table contans 256 rows inspite of just 1 record)
I checked the OBI guide and found that parameters $$Start_date and $$End_Date needs to be set.
I am not sure if Analysis end date in DAC-> ETL prefernces is same as $$End_Date . Also, I could not see any start date parameter. I have checked mapping parameter as well.
It would be really great if you could tell the exact steps followed to extend the datawarehouse.

Time Dimension Wizard Hangs Creating Map

I'm running OWB 11.1.0.7 and trying to create a time dimension using the time dimension wizard and when I get to the progress screen it just hangs when it gets to 60% progress and the message window says it's creating the map.
There is no indication of any kind of error that might say why it's hanging, or any indication other than "creating map" that says what it's trying to do at the time. Is there anyway to dig under the covers anywhere to see what the wizard is trying to do that is causing it to hang?
Another question is how I can create the same kind of dimension manually that the wizard would have created? It appears to be creating a special kind of dimension, specifically for time but I can't see anyway to create it manually if the wizard doesn't work. I found another message that mentioned a time dimension template that the wizard uses but that post was about how to use OMBPlus to remove the time template type so it could be edited manually, not how to create one using it.
So, anyone have any suggestions for how I can track this problem down and/or create that time dimension manually? Thanks.
Regards, Bob

Are you using OWB 11.1.0.7 or greater?
We encountered the same problem from that version on. try to use an older version (11.1.0.6 client definitely works).
We will put this into metalink as it is definitely a bug.
Cheers,
Knut

Time Dimension _ Report on Current month and previous year.TOTAL

Hi,
My Time dimension is as under .
July2013, Aug2013 ,......Jan2014,,,. June2014 = 2014.TOTAL
My requirement is to report based on context view along with previous period total (.TOTAL) values.
For example : If i am using CV as : 2013.JUL my report should show 2013.JUL & 2013.TOTAL
If i am using CV as : 2014.JUN my report should show 2014.JUN & 2013.TOTAL
Can anyone provide some direction to build the above scenario
Regards
Krishna

Hi,
do you have the property YEAR in your time dimension filled out? If so, you could use EPMDIMENSIONOVERRIDE function together with an Excel cell.
Let me build an example:
Assuming your report has two columns, one for the current month and one for the total year. I would split this into two, one report per month, i.e. column. The first report will show the current period, the second the total. Somewhere in the header of your sheet use:
EPMDimensionOverride("001";"TIME";EPMMemberProperty(;EPMContextMember(;"TIME");"YEAR")&".TOTAL")
001 is the default ID of the second report.
BR,
Arnold

Weeks and Months roll-up in Time dimension

We need to change our Time dimension roll-up ('Total Year' 'Quarter' 'Month') as some users want to see the week level as well. What is the best way to do that? I still want to use the DTS and both need to be Time dimension. Thanks for your reply.

Assuming you have data down to the daily level, another option would be to create an alternate hierarchy in your time dimension. You would use your level zero members (individual days) that are currently rolling up into months and share them. Your Time dimension you would then have two children; "By Months" and "By Weeks".Under months you would have Jan, Feb, Mar, etc. Under Weeks you would have Week 1, Week 2, ... Week 52. The level zero members are the same for both. What I like about this method is it does not add any overhead since you are using shared members, and you don't run into a situation where your total "by months" varies from your total "by weeks". Since your base level data is the same you can have confidence that your data will tie out. Of course the end user needs to decide when they run their report wether they want to run by months or by weeks. My users were more than happy to do this when I explained there was no practical way to have months and weeks in the same hierarchy since weeks will overlap months. I have an outline set this way and I am using DTS effectively. If you want, send me an email and I will send you an OTL file with the Time dimension as I explained.Best Regards;

Fiscal Year Time Dimension - Month Time Span

I have a need to create a fiscal year time dimension and I created using Time Wizard in OWB. When I was developing BI Reports, I found some inconsistencies in Time dimension as it was showing figures in all the months of my first quarter of new financial year that Apr 2009 – Mar 2010. On close inspection of Time Dimension table I found that fiscal_month_time_span value varies from 35 to 28 to 29 for different months. i.e. it’s not matching with corresponding Calendar month. This means I am unable to plot data correctly using fiscal year dimension.
Maybe this is how it works in Oracle, but this does not solve my problem. Could anyone give me some solution for this ? My Fiscal year is from ‘1- Apr – 09’ to ’31-Mar-10’. In my reports I would like to see transactions for April month when I use month attribute of time dimension like it happens in Calendar time dimension.

May be someone from
Forum: Business Intelligence Suite Enterprise Edition can answer this .
Please post it there.

Date and Time dimensions

After reading the following article, I have decided to use SSAS dimension wizard for generating our Date dimension, which creates a DATETIME PK.
http://www.made2mentor.com/2011/05/date-vs-integer-datatypes-as-primary-key-for-date-dimensions/
I have also created a separate Time dimension as granularity of an hour is required.
The Time dimension is very simple and only contains a surrogate key (INTEGER) and actual time in hours (VARCHAR).
DimTime(TimeKey, TimeInHours)
Our Fact table will now have a link to both the Date and Time dimension using the PK's.
Our analysis is required by hour, day, week, month and year.
My query is; Will this current structure cause any problems when creating MDX scripts to analyse our data (i.e. drilldown and rollup queries) Hour - Day - Week - Month - Year

Hi Darren,
According to your description, there a day and hour granularity in your fact table, so you want to a hierarchy like Hour - Day - Week - Month - Year, right?
In your scenario, you created a time table that only contains a surrogate key (INTEGER) and actual time in hours (VARCHAR). We cannot create a Hour - Day - Week - Month - Year hierarchy without ant relationship between date table and time table. As per my understanding,
you need create a foreigner key in time table, and join those table in the data source view, then you can create such a hierarchy. Here are some links about create time dimension, please see:
http://www.ssas-info.com/analysis-services-articles/59-time-dimension/1224-date-and-time-dimensions-template
http://www.codeproject.com/Articles/25852/Creating-Time-Dimension-in-Microsoft-Analysis-Serv
Regards,
Charlie Liao
TechNet Community Support

Time Dimension- Wk and Month

Similar Messages

Maybe you are looking for