MSTR Inventory metrics or No Aggregate metrics

TN20363: How do non-aggregatable metrics work in MicroStrategy SQL Generation Engine 9.x?

https://success.microstrategy.com/t5/Architect/TN20363-How-do-non-aggregatable-metrics-work-in-MicroStrategy/ta-p/180533

Some kinds of fact data are not valid for aggregation with respect to one or more dimensions. The classic example is inventory with respect to time. The fact that there were 300 units in inventory yesterday is not relevant to the 200 units in inventory today. Adding those values together results in an inflated total, 500, that does not reflect the actual size of the inventory. What is relevant, rather than the sum, is the last atomic inventory value in a given period of time.

In MicroStrategy SQL Generation Engine 9.x, non-aggregatable metrics are used for this kind of calculation. Non-aggregatable metrics are defined using the dimensionality section of the metric editor. The "Grouping" property has six possible settings, of which four specify non-aggregating behavior:

Beginning (fact)
Ending (fact)
Beginning (lookup)
Ending (lookup)

"Fact" and "lookup" define how the first or last atomic value is identified within the period. "Fact" means to use the first/last existing value in the fact table, while "lookup" goes to the first or last child element of the element to be displayed on the report. For example, a fact table might contain data such as the following:

|| Quarter_ID

Month_ID	Fact
20061	200601	10
20061	200602	20
20061	200603	30
20062	200604	40
20062	200605	50

No data exist for June 2006. An ending fact metric would return 30 for Q1 2006 and 50 for Q2 2006; the corresponding ending lookup metric would return 30 for Q1 2006, but the value for Q2 2006 would be null. This is because the lookup table identifies June 2006 as the last month in the second quarter, but the fact table does not have any data for that month.

Dimensionality vs. fact table attributes
The actual calculation level of a non-aggretable metric will be the lowest level attribute from the non-aggregatable hierarchy that is present on the fact table. If the fact table includes the Day attribute, day-level values would be the result. If, instead, Month is the lowest fact table level, then the metric would produce month-level values.The level of the first or last "x" within "y" depends on the granularity of the fact table.

If fact tables exist at multiple levels within a non-aggregatable hierarchy, the attribute chosen as the dimensionality target determines the calculation level by influencing the selection of the fact table. If Month is the target attribute, the Engine will never choose a fact table at a higher level than Month (never Quarter or Year). Thus the first or last month-level value will be presented in the report.

Note: The final determination of the calculation level is based on the fact table. The Engine uses normal fact table selection logic for non-aggregatable metrics, taking into account dimensionality and filtering attributes. The Engine will revert to a lower-level fact table if that is the only way to support all of the grouping and filtering attributes. For example, if a metric indicates non-aggregatable dimensionality at the level of Month, but a report includes attributes that require a day-level fact table, the day-level table will be used. Even though the metric specified Month level, the report will display the first or last day's value within the time period because the fact table demands it.

Beginning/ending (fact) dimensionality
Beginning or ending (fact) metrics calculate according to the following procedure:

Calculate the metric, including both the dimensional level attribute ("parent") and the fact table level attribute ("child"). Including the fact table level attribute at this stage ensures that the non-aggregatable hierarchy will not be subject to aggregation.
For each "parent" element, find the first or last "child" element that exists in step 1.
The metric's final result is the intersection of steps 1 and 2.

For example, an "End on Hand" metric may be defined with the following dimensionality. Month is used as the dimensional attribute because the fact exists at the level of Month.

external image TN5200-8X-2457_1.gif

A report calculating this metric over Quarters generates the following SQL. (Note that a filter restricts the results to one quarter. This is because the MicroStrategy Tutorial project uses a partitioned fact table for the End on Hand fact. The filter reduces the size of the SQL and makes a better illustration.)

Partitioning pre-query -- this is independent of non-aggregatable metrics

select distinct a11.PBTNAME PBTNAME

from PMT_INVENTORY a11

where a11.QUARTER_ID in (20074)

Month-level values, with Quarter ID included

select a12.QUARTER_ID QUARTER_ID,

a11.MONTH_ID MONTH_ID,

sum(a11.EOH_QTY) WJXBFS1

into #ZZTTS0200F6PO000

from INVENTORY_Q4_2007 a11

join LU_MONTH a12

on (a11.MONTH_ID = a12.MONTH_ID)

where a12.QUARTER_ID in (20074)

group by a12.QUARTER_ID,

a11.MONTH_ID

What is the last Month found in the metric table for each Quarter?

select pc11.QUARTER_ID QUARTER_ID,

max(pc11.MONTH_ID) WJXBFS1

into #ZZTTS0200F6MB001

from #ZZTTS0200F6PO000 pc11

group by pc11.QUARTER_ID

Intersect (inner join) the tables to preserve only the last Month's value from each Quarter

select distinct pa11.QUARTER_ID QUARTER_ID,

a13.QUARTER_DESC QUARTER_DESC,

pa11.WJXBFS1 WJXBFS1

from #ZZTTS0200F6PO000 pa11

join #ZZTTS0200F6MB001 pa12

on (pa11.MONTH_ID = pa12.WJXBFS1 and

pa11.QUARTER_ID = pa12.QUARTER_ID)

join LU_QUARTER a13

on (pa11.QUARTER_ID = a13.QUARTER_ID)

drop table #ZZTTS0200F6PO000

drop table #ZZTTS0200F6MB001
Beginning/ending (lookup) dimensionality
Beginning or ending (lookup) metrics follow a slightly different procedure.

Create a table containing the first or last child elements (fact table level) corresponding on the parent elements to be displayed (dimensional level). Note that there is no need to include the parent attribute in the SELECT clause of this pass, because the pass will be used only for filtering the fact table. (The parent attribute is in the GROUP BY clause.)
Calculate the metric, grouping by the dimensional level attribute ("parent") but filtering according to the child elements from step 1.

Determine the latest Months in each Quarter, according to lookup table

select max(c11.MONTH_ID) MONTH_ID

into #ZZTTS0200GZOP000

from LU_MONTH c11

where c11.QUARTER_ID in (20074)

group by c11.QUARTER_ID

Partitioning pre-query -- this is independent of non-aggregatable metrics

select distinct a11.PBTNAME PBTNAME

from PMT_INVENTORY a11

where a11.QUARTER_ID in (20074)

// Calculate the metric, filtering on the set of last months in their quarters

select a13.QUARTER_ID QUARTER_ID,

max(a14.QUARTER_DESC) QUARTER_DESC,

sum(a11.EOH_QTY) WJXBFS1

from INVENTORY_Q4_2007 a11

join #ZZTTS0200GZOP000 pa12

on (a11.MONTH_ID = pa12.MONTH_ID)

join LU_MONTH a13

on (a11.MONTH_ID = a13.MONTH_ID)

join LU_QUARTER a14

on (a13.QUARTER_ID = a14.QUARTER_ID)

where a13.QUARTER_ID in (20074)

group by a13.QUARTER_ID

drop table #ZZTTS0200GZOP000

Comments

Microstrategy Custom number formatting symbols

Custom number formatting symbols If none of the built-in number formats meet your needs, you can create your own custom format in the Number tab of the Format Cells dialog box. Select Custom as the Category and create the format using the number format symbols listed in the table below. Each custom format can have up to four optional sections, one each for: Positive numbers Negative numbers Zeros Text Each section is optional. Separate the sections by semicolons, as shown in the example below: #,###;(#,###);0;"Error: Entry must be numeric" For more examples, see Custom number formatting examples . To jump to a section of the formatting symbol table, click one of the following: Numeric symbols Character/text symbols Date and time symbols Text color symbols Currency symbols Conditional symbols Numeric symbols For details on how numeric symbols apply to the Big Decimal data type, refer to the Project Design Guide . ...

Slowly Changing Dimension Types 0, 4, 5, 6 and 7

Slowly Changing Dimension Types 0, 4, 5, 6 and 7 Ralph introduced the concept of slowly changing dimension (SCD) attributes in 1996. Dimensional modelers, in conjunction with the business’s data governance representatives, must specify the data warehouse’s response to operational attribute value changes. Core SCD approaches: Type 1 (overwrite), Type 2 (add a row), and Type 3 (add a column). Since legibility is a key component of the Kimball mantra, we sometimes wish Ralph had given these techniques more descriptive names, such as “overwrite” instead of “type 1.” But at this point, the SCD type numbers are part of our industry’s vernacular. We have written about more advanced SCD patterns, such as the 2005 article entitled “ Slowly Changing Dimensions are Not Always as Easy as 1, 2, 3. ” However, we’ve not consistently named the more advanced and hybrid techniques. With the third edition of The Data Warehouse Toolkit (Wiley, 201...

Internationalization Design Technics

Microstrategy Internationalization Design Technics MicroStrategy supports data internationalization through two different techniques. You can either provide translated data through the use of extra tables and columns, or you can provide separate databases to store your translated data. These techniques are described below: You can support data internationalization in your database by using separate tables and columns to store your translated data. You can use various combinations of tables and columns to support and identify the translated data in your database. To support displaying the name of each month in multiple languages, you can include the translated names in a separate column, one for each required language, within the same table. Each column can use a suffix to identify that the column contains translated data for a certain language. The same LU_MONTH_OF_YEAR table with translated data for the Spanish and German langua...

Custom Subtotal Displays in MicroStrategy

Defining custom subtotal displays in MicroStrategy By default, when users apply subtotals in a report, the name of the subtotal is displayed in the subtotal line items that appear in the report. Users can use custom subtotals to give more control over the characteristics of a subtotal. Custom subtotals allow users to define custom subtotal line items that appear on the reports U sers can make the subtotal name dynamic by typing special characters in the subtotal name field as listed in the following table. Character Description #A The name of the attribute under which the subtotal appears. #P The name of the attribute to the left of, or above the attribute under which the subtotal appears. #0 All the forms of the parent element. #1 The first form of the parent element reading from left to right or from top to bottom. #2 The second form of the parent element reading from left to right or from top to bottom. #3 The third form of th...

Case functions Microstrategy

Ca se functions Microstrategy Case functions return specified data in a SQL query based on the evaluation of user-defined conditions. In general, a user specifies a list of conditions and corresponding return values. Case This function evaluates multiple expressions until a condition is determined to be true, then returns a corresponding value. If all conditions are false, a default value is returned. Case can be used for categorizing data based on multiple conditions. This is a single-value function. Syntax Case ( Condition1 , ReturnValue1 , Condition2 , ReturnValue2 ,..., DefaultValue ) Example Case(([Total Revenue] < 300000), 0, ([Total Revenue] < 600000), 1, 2) sum(Case (Day@DESC in (“Sat”,”Sun”), Sales, 0) {~+} Sum(Case(Category@DESC In("Books","Electronics"),Revenue,0)){~+} CaseV (case vector) CaseV evaluates a single metric and returns different values according to the results. It can be used to perfo...

Configure a report for use with Bulk Export in MicroStrategy

Configure a report for use with Bulk Export in MicroStrategy The Bulk Export feature enables a large report to be saved as a delimited text file. Using this feature, it is possible to retrieve result sets from a large dataset without having to load the entire dataset into memory. PS: Once a report is setup for bulk export it cannot be used as a regular report. So if the report needs to be run as a normal report and as a bulk export report, the first step is to make a copy of the report for use with bulk export. Configure Bulk Export Bulk Export options are only available in MicroStrategy Developer. Open a 3-tier connection using MicroStrategy Developer and edit the desired report. Go to 'Data' on the top menu bar. Select 'Configure Bulk Export': Specify any additional desired configuration options. General Settings Bulk export database instance : This is the database instance to use to store the bulk export results. Temporary tables w...

Configure Connection Mapping in Microstrategy

Configure Connection Mapping in Microstrategy The following steps demonstrate the second scenario where two different data warehouses are used within the same project: Create two different database connections - One that points to the data warehouse for the European users and the other that points to the data warehouse for USA users as shown below: Select Europe as the default database connection for the database Instance as seen below: Go to P...

Sort by metric subtotals and attribute elements together in Microstrategy

Sort by metric subtotals and attribute elements together in Microstrategy Users may observer that when creating a report that contains advance sorting with a metric that contains subtotals the report results appear to be only sorted by the metric values specified. Even if a sort is specified for the attribute elements on the report, the results in the report appear as if the attribute sort was not defined. In the screenshot below, the results for a report are shown where the Advance Sorting option 'Sort metrics hierarchically using: Total' is selected. For this report, a second sort is defined on the Customer Gender - 'DESC' form, users would notice that the ordering of the this attribute is not consistent: The sort definition for the report is shown below: CAUSE: When the option to 'Sort metrics hierarchically using: Total' option is selected, the MicroStrategy Engine first sorts the results based on the Total values, and then sorts th...

Star schemas and aggregate (or summary) fact tables Aggregate tables can further improve query performance by reducing the number of rows over which higher-level metrics must be aggregated. However, the use of aggregate tables with dimension tables is not a valid physical modeling strategy. Whenever aggregation is performed over fact data, it is a general requirement that tables joined to the fact table must be at the same attribute level or at a higher level. If the auxiliary table is at a lower level, fact rows will be replicated prior to aggregation and this will result in inflated metric values (also known as "multiple counting"). With the above Time dimension table, a fact table at the level of Day functions correctly because there is exactly one row in DIM_TIME for each day. To aggregate the facts to the level of Quarter, it is valid to join the fact table to the dimension table and group by the quarter ID from the dimension table. Sql select DT...

The logical table size calculation in Microstrategy

The logical table size calculation in Microstrategy The logical table size is an integer number that represents the granularity or level of aggregation of a particular table. It is called 'logical' because it is not related to the physical size of the tables (number of rows). It is calculated according to the attribute IDs that are present in the table and their level in the system hierarchy. Even though, the number does not reveal the actual number of rows in the table, it is an accurate way of measuring a table size without having to access its contents. MicroStrategy Engine utilizes an algorithm based on attribute keys to calculate the Logical Table Size (LTS): Given the following tables: The algorithm that calculates the table sizes performs the following steps: Calculate the number of levels per hierarchy: Hierarchy 1: 3 Hierarchy 2: 4 Calculate each attribute individual weight according to the level in the hierarchy (level in hierarchy/number of ...

Microstrategy Analytics

Search This Blog