Aggregation Behavior, Something To Do With MembersWithData Setting But Can't Nail It Down
May 22, 2008
I have a dimension with a hierarchy. In this hierarchy it is possible for a non leaf member to get loaded with a fact measure.
Using a sales organization as an example, let's say
(1) Bill and Ted are sales reps, and they both report to John.
(2) John, not only manages Bill and Ted's sales, but he also makes sales of his own.
(3) Bill sells 10 units, Ted sells 8 units, and John sells 5 units.
In this example, John is the non-leaf member because he is above Bill and Ted in the hierarchy. When this data is aggregated in the hierarchy, we want to make sure that 10 for Bill plus 8 for Ted do not overwrite the 5 for John. Instead, we want to include the 5 for John with Bill and Ted's total of 18, thus getting a total of 23 for John.
Here is my question. What is the default aggregation behavior? The behavior I get is that the non leaf data is overwritten by the sum of the leaf data. So in the example above, I get 18 for John, not 23.
Now, from my research, I think the MembersWithData setting on a dimension attribute as something to do with this, but I can't nail this down.
we're trying to get a better understanding of how RS behaves when parameters are being set. We see quirky behavior that is a little difficult to describe. Right now we assume that if the revolving green circle (with the phrase "Report is being generated" beneath it) doesnt appear, the report really wasnt rendered properly, even if the report region changes.
One peculiarity that seems pretty consistent is on reports we've prototyped with "from" and "to" date parameters. It seems that when we set one date (doesnt matter which is 1st) things progress normally, ie no "report clearing event" occurs as a result of setting cursor focus in the calendar control and changing its value. The report region doesnt change from what showed previously. But trying to set focus on 2nd (doesnt matter if its "from" date or "to" date, just that its the 2nd date being set) always seems to trigger some kind of event that 1) doesnt allow focus to be on that text box, 2) blanks out the report region including headings. Only after this "event" occurs, can we set focus on the 2nd date, change the value and click the "view report" button for rerendering.
We see similar types of behavior with other types of parameters that include multi value dropdowns and booleans. The toughest part of this is trying to explain it to our users. On some parameters, the event always occurs every time they are changed. On other parameters, it appears that the event only occurs if another parameter was changed beforehand.
I believe we've even seen headings with no data rendered, thinking temporarily that no rows were returned, just to find out that by clicking the "view report" button there really was data to be reported based on current filters. Unfortunately I cant reproduce this scenario when I want to.
I would like to know what happens when a very large reference data set for a lookup transform with full caching enabled is getting loaded during package execution and the computer memory runs out or is very low. Does SSIS a) give an out of memory error of some sort b) resort to a no caching or partial caching mode c) maintain the full caching mode but will switch to using the paging file(virtual memory).
I think it will resort to using the page file in which case the benefits of in memory lookups are lost and performance would suffer. If I cannot upgrade the memory or shrink the reference set somehow, i should switch that lookup task to use partial caching or no caching with an indexed lookup table. Would this make sense?
I ran into a variety of problems trying to set a script task breakpoint in a package containing multiple script tasks. The debugger apparently treats the breakpoint as if it were set in ALL tasks in the package, not just the one in which it is actually set.
At best this results in hitting breakpoints in scripts where they are not set and at worst the debugger brings up the "Send error report" dialog and quits (while the package continues to run). The latter seems to happen most often when the script with the breakpoint has more lines than an earlier script and the breakpoint is set at a line number that exceeds the number of lines in the earlier script--it bombs when the earlier, shorter script starts.
To get the debugger to work under these circumstances I had to add some nonsense code like
While False Dim i as Integer = 0 End While to every script, at the same line number near the beginning of the script (line 40, for example). I then set a breakpoint on the middle statement in one of the scripts (doesn't matter which) to cause the debugger to open at runtime. It doesn't hit the breakpoint because the line is never executed. If the breakpoint is set on a line that can be executed in any script in the package, bad things tend to happen. I then add a "stop" statement to the script that I want to debug. This only works if the debugger is already open, hence the dummy breakpoint above. This workaround is usable, but I am debugging a package that has quite a few scripts and having to insert dummy code in all of them at a fixed line number is rather inconvenient. I would really like to see breakpoints work the way one would expect--only in the scripts where they are set. Is there some other, easier way around this problem? Is there at least an easier way to get the debugger to open so that "stop" will work?
I'm using an Aggregation task to summarize an input file by item and week before inserting it into a SQL table.
Two of the fields I'm summing, because their totals per record can occasionally exceed 32k, are defined as int (I4) instead of smallint (I2). However, the summarized total never exceeds the value an int can hold.
I ran into a problem on the insert, however, with SSIS telling me it couldn't insert an I8 value into an I4 table field. I discovered the metadata for the summed totals had automatically been set to bigint (I8), and the mapping was failing.
I didn't see a way to change that metadata within the Aggregation task itself, so I added a Data Conversion task to convert the totals to four-byte signed integers and enable the mapping. Was that the proper workaround?
Hi. i have a problem that related to aggregation functions. i have a table (requests)that consist with the following fields:event_id, request_type,data. i want to present a report that for each event_id it will show the number of records where the request_type value is:"VMR".
i wrote the following query- SELECT event_id, Count(request_type) FROM requests WHERE ((requestType)="VMR") GROUP BY event_id
the problem is that:if event_id don't has a record with the value "VMR" it will not be showen in the report and the goal is to present those kind of events with the value 0. who can i make it happen.
Can anyone help me with an problem I have come across in my databasedesign.I have a primary table and a related table with 3 child records (eachwith a numeric field). I require a query to return the primary keyfrom the main table and the PRODUCT (i.e. all numeric valuesmultiplied together) of the three child records, much like a SUM wouldadd them together.Any help would be gratefully received!Tony.
I am new to the reporting services and I've been working on problem in one of my reports all day long and after 8 hours of frusturation I decided ask for a profesional help.
Ok here is my problem: I have a report that calculates the amount of meetings with our clients. The dataset contains an activity_id field that we assign for each our meetings with our clients. SSRS counts these meetings and shows it in a drilldown enabled report. Everything seems fine on the report except that someof the activities involves few different clients and SSRS is not counting the activities multiple times in region drilldown as there is only one activity id associates in that region even though it contains different companies. And I want those companies to be calculated in too.
From the crude drawing below I wanted to explain my dilemma visually. As it can be seen the total number of meetings we had is actually 40. But as we had 3 activities that involves more than 1 clients it only gives 37 as a count. I would like to know is there a way to make the report count the same activity multiple times if activity_id is associated with more than one clients.
I hope I managed to explain my problem
********************************************************************************************** Manager Region Market Company Meeting Detail
I know we can design Aggregation usage for a dimension within a cube. For example, you can choose Full , None, Unrestricted and Default method. I assume this is the Global settings.
Is it possible that Ican do the same thing for the attributes within a dimension? I cannot find such property in attributes. Does this means we can't control which attributes we want Analysis Services to aggregate in a dimension?
I have the following three tables : Account (Id int, AccountName nvarchar(25)) Role (id int, Rights int) AccountRole (AccountID, RoleID)
In Role table - Rights Column is a bit map where in each bit would refer to access to a method. One account can be associated with multiple roles - AccountRole table is used for representing the N:N relation.
I want to develop a store procedure - which would return all AccountName and their Consolidated Rights. Basically I want to do a BitWise OR operation for all the Rights in the Aggregation instead of the SUM as shown in the following statement.
SELECT Account.Name, SUM(Role.Rights) FROM Account WITH (NOLOCK) JOIN RoleAccount ON RoleAccount.AccountID = Account.Id JOIN Role ON RoleAccount.RoleId = Role.Id GROUP BY Account.Name
I have a table imported from a legacy Oracle database that stores values vertically in name/value pairs. I store it in table-type variable that is an exact copy of the structure:
If this information were pivoted horizontally: OMNI_NUMBER would be the primary key.
TIMESTAMP is a 10-digit integer that represents the number of seconds since 1/1/1970 UTC that requires additional conversion. DATA_TYPE is not the data type. It is a general categorization of the next two columns.PARAMTER would be the column headings if it were horizontal..PARAMETER_VALUE would be the data value in that column.
I would like to try to use PIVOT to list the PARAMETER column values as column headers. This seems to work fine. What's confusing me is that I'd like it to list the PARAMETER_VALUE column values as raw data, just as it is in the source version, without having to apply some sort of aggregate function to it. Here's a CSV sample of the data you can paste into Excel. I'm trying to transform this:
I have a Sales cube and I want to be able to display Products and the Price at which they were purchased in a given period. I have a Product dimension while Price is a measure in my Sales Fact Table. Is there a way to have a "Group By" aggregation instead of a Sum? This way, I can show Products grouped by their list price.
My Product Dimension consists of product_numbers (such as 100, 101, 102 etc...) My Sales Fact Table consists of sales data (such as product_key, price, net_sales, returns, etc...) at the transaction level.
where Net_sales and Returns are "summed" and Price is simply a "Group By". In other words, this report would show the net sales and returns by product for a given price.
I'd rather not use a Price dimension since we have hundreds of products at hundreds of different prices. Moreover, this data is already in the Sales Fact table.
I am trying to port a database from SQL Server 2005 Express to SQLServer 2005 Standard. Idetached my database from the SQL Sever 2005 Express instance andattached it to the SQL Server 2005 Standard instance. Whenever I do aselect on tables in the database I have no problem. However if Iattempt to open or modify a table by selecting open or modify from thepopup menu, I get the following error --TITLE: Microsoft SQL Server Management Studio------------------------------Class does not support aggregation (or class object is remote)(Exception from HRESULT: 0x80040110 (CLASS_E_NOAGGREGATION))(Microsoft.SqlServer.SqlTools.VSIntegration)------------------------------BUTTONS:OK------------------------------I tried repairing the .NET 2.0 framework but to no avail.Please help!!!Thanks,
Hello, I have a resultset as follows:fields: Name, RankIDvalues:Prorduct A, 4Product B, 33Product C, 221(etc)Name is always unique. RankID may not be.I want to take that result set and basically pivot it to have the Namevalues as columns, and the RankID of each one as the data. So youwould end up with only one row like:Product A | Product B | Product C | etc4 | 33 | 221 | etcIs this possible? I do not want to sum the data or anything, simplyrotate it sort of.Any advice is appreciated.
I need to create a report containing the headcounts of the employees for each month. The data I can retrieve from the database looks like this:
Name StartEmp EndEmp
tom 1/jan/05 15/mrt/05
dirk 1/jan/05 31/mrt/06
jan 1/feb/05 NULL
In order to get the right information for the report the information should be represented in the following way:
January 2005 February 2005 March 2005 "April" 2005
Tom 1 1 0,5 1
Dirk 1 1 1 1
Jan 0 1 1 1
HeadCount 2 3 2,5 3 (there should be more data in the 2nd table but it's a sliced representation of the data. Sliced by 2 report parameters: startdate (=1 January 2005) and enddate (=30 April 2005) )
I need some help about building up the query to aggregate the info from table1 to the format of table2.
my package is failing every now and then on the following error:
SSIS Error Code DTS_E_OLEDBERROR. An OLE DB error has occurred. Error code: 0x80004005. An OLE DB record is available. Source: "Microsoft SQL Native Client" Hresult: 0x00040EDA Description: "Warning: Null value is eliminated by an aggregate or other SET operation.". yes, I do have a null aggregation, but in all my tests this didn't cause a failure... anyone had experience with this?
I have daily data that I wish to summarise into monthly. The field [Close] will return the last day of the month that has a value.The query below summarises the data as I want it but using the max of [Close] in the month instead rather than the Last value for the month. In Access I used to use Last function that worked perfectly. From googling, Last doesn't seem to exist in T-SQL but there must be a way to do it.
D.YearMonth_Number,max(AT.[Close]) FROM tbl_Date ASD INNERJOINtbl_ASXStats ASAT OND.YearMonthDay_Number =AT.YearMonthDay_Number
Is this possible in any way? Basically I have a standard pair of tables in a master/detail relationship. I need to somehow be able to query the child table for all children related to a parent record and return some sort of aggregation of a certain column. The problem is that I still need to know what each value was, so it can't be something like a SUM or an AVG.
As if that wasn't complicated enough, this will be done on 1000+ records at a time and is being used in a binding situation.
I've figured out how to do this with standard SQL Server with some dirty FOR XML hacks, but I'm beginning to think that it's not in any way possible with SQL Compact.
I have a multi dimensional cube. Among the dimensions I have, there is one dimension that has one hierarchy defined. When I view any of my measures on this hierarchy, I don't want the measures to be aggregated. I read some threads here having to do with semi additive behaviors, but Im not sure if it applies to my case. (or maybe it does and I'm just not getting it)
The purpose of "AnotherPersonID" is to have a self join that describes a certain relationship. So, in this case, Person B and C are related to Person A. I describe this relationship using a hierarchy where A is a parent of B and C. However, I don't want the measures for Person_A to be replaced by the sum of the measures in Person_B and Person_C.
Why do I have a hierarchy if I don't want to sum the numbers? This is motivated by the need to dynamically report on A's numbers when B or C's numbers are reported. I'm just using persons as an example, but, in my case I have a bunch of members that can be paired with another member in that dimension. And when I report on a member that joins to another member in that dimension, I need to dynamically report on that member's measures as well.
Lastly, If the use of a hierarchy is not the best approach for this, would you recommend another approach?
Hi, I have following data coming from prior transformations.
Id Unit Name NewFlag ___ __________ ____________
1 JacuzziBox 1 2 Hummer H2 0 3 Waste dumper 0
and so on.
I want aggregate task to result 2 fields (without any GROUP BY-ing), so that one contains total number of rows or records, and one contain total number of records with NewFlag=1.
Just wondering if it is possible by using Aggregate and how.
Class does not support aggregation (or class object is remote) (Exception from HRESULT: 0x80040110 (CLASS_E_NOAGGREGATION)) (Microsoft.SqlServer.SqlTools.VSIntegration)
Program Location:
at Microsoft.SqlServer.Management.UI.VSIntegration.ISqlToolsVSHlpFactory.CreateHierarchyWrapper(Guid& clsidInnerHier) at Microsoft.SqlServer.Management.UI.VSIntegration.SqlWorkbenchHierarchy.CreateHierarchyWrapper(Type typeToWrap, ILocalRegistry localReg) at Microsoft.SqlServer.Management.UI.VSIntegration.Editors.ScriptFactory.get_SqlVirtualProject() at Microsoft.SqlServer.Management.UI.VSIntegration.Editors.VsDocumentMenuItem.InvokeDesigner(IManagedConnection connection) at Microsoft.SqlServer.Management.UI.VSIntegration.Editors.VsDocumentMenuItem.Invoke()Hi there,
I'm using SQL Server 2005 Dev. with Windows XP Pro. SP2. 50GB free space with 1GB of RAM After I installed SQL Server with no problems. I tried to open a table in AdventureWorks database and I'm getting an error:
I have a problem: I have various cube packed in a virtual cube to give me distinctCount Values of customers who purchased products (DC customer] and purchased products (DC product) along with Saleries [RW] over numerous dimensions such as productgroups customergroups etc. In the calculated measure [RW kum] I can sum the the salaries from the beginning of the year until [time].currentmember with the following MDX-Code: With member [Measures].[RW kum] As 'Sum(Periodstodate([time].[year]).[Measures].[RW])' Select {[Measures].[RW],[Measures].[RW kum], [Measures].[DC customer]} on rows, [Productgroup].[Itemgroup].Allmembers on columns From [vcub_Sales] Where [Time].[2005].[march]
IF I use this PeriodsToDate-Formular with the the measure DC customer I will get the sum of distinctcounts of the periods.
What I want is the distinct count of all customers who have purchased since the beginning of the Year until the [Time].currentmember in the same Level.
Given this problem, do I.. 1. have make another Cube or 2. another time dimension or 3. is there a way in MDX to get the result?
eventually I started using DERIVED Tables and Sub queries within them if needed. I like choking down all the queries in the selects joining them and having your result set select there to choose from all the aliases, it also resolves while doing this. So much easier IMO than using CTE's or TEMP Tables. I was big on temp tables for a while...
I'm curious though if you want to count a type of criteria in a column do you use a standard case in your inner query choking it down just for those particular counts? Then do another case on the other criteria and END AS with the alias name?
E.G.
CASE WHEN COLUMN1 = PIZZA THEN COUNT(COLUMN1) END AS PIZZACOUNT CASE WHEN COLUMN1 = ROOTBEER THEN COUNT(COLUMN!) END AS ROOTBEERCOUNT ETC...........
I have a problem that requires me to write Daily sum aggregation script, I wrote the script; But what I'm not sure of is the "Daily" part, my script will give me only give me results when I execute it. How to make this code daily?
SELECT Receipt ,"Date" ,Item ,Reason ,Division ,SUM(Cost) AS Cost
I have a problem that requires me to write Daily sum aggregation script, I wrote the script; But what I'm not sure of is the "Daily" part, my script will give me only give me results when I execute it.
I manage a legacy system that dumps it's data into a number of different databases (same schema) on a nightly basis using bulk insert. I need to formulate a strategy for efficiently aggregating that data into a single database right after these nightly extractions complete. Here is my current stategy:
1. Duplicate the legacy system's database schema and add an identifier column to specify which database the data loaded from.
2. Each night, delete all records in the table.
3. Each night, for each database:
3a. Set each table's default value to a value that references the current database being loaded.
3b. Use the legacy system's flat files and format files to bulk insert into the database.
3c. Clear the default value.
What other steps would faciliate performance? Dropping and recreating the indexes? Does anyone forsee faults in this strategy?
I have 3 partitions using a year grouping. Current year, previous 4 years, older than 5 years. I have two measure groups, one is a distinct count, so I actually have 6 partitions.I also use usage based optimization to build my aggregations. Should each partition have a separate aggregation or should there be one for each measure group?
