1

Topic: Error at handling of cube Analysis services 2008R2 standard

Colleagues, help. At cube handling there is an error: Errors of the unit of storage OLAP: the attribute key is not found at handling: the table: "dbo_IMSFact_forCube", a column: "PRODUCTID", value: "80174". Attribute "Pack Key".
I checked in measurement "Products" this key is, in the fact table too is. If to eliminate this Pack Key from the table of the facts an error it is saved, but swears already on another PRODUCTID which too is both in the fact table and in measurement.
What do I do not so?

2

Re: Error at handling of cube Analysis services 2008R2 standard

ondorsal;
H'm, I would make the following:
1.  at first measurements. At processing of measurement of products - looked in processing of attribute/hierarchy Pack Key, looked at request which fulfills SSAS,   the request in SSMS and checked up about presence in it of it Pack Key.
2.  groups of measures, there by errors - the request similarly derived, looked at it. If it is necessary - launched in SSMS, found the specified line and checked up that there not so.
The piece in that SSAS for the measurements based on several tables, does in between join - and it can be both left and inner. As a result of a line which like as can eat not to enter into sampling for attribute.

3

Re: Error at handling of cube Analysis services 2008R2 standard

Ferdipux;
Daw materialize at these measurements remove and there will be to you a happiness

4

Re: Error at handling of cube Analysis services 2008R2 standard

Ferdipux wrote:

ondorsal;
H'm, I would make the following:
1.  at first measurements. At processing of measurement of products - looked in processing of attribute/hierarchy Pack Key, looked at request which fulfills SSAS,   the request in SSMS and checked up about presence in it of it Pack Key.
2.  groups of measures, there by errors - the request similarly derived, looked at it. If it is necessary - launched in SSMS, found the specified line and checked up that there not so.
The piece in that SSAS for the measurements based on several tables, does in between join - and it can be both left and inner. As a result of a line which like as can eat not to enter into sampling for attribute.

Thanks for the answer.
1). As at processing of measurement of products - to look in processing of attribute/hierarchy Pack Key, to look at request which fulfills SSAS,  request in SSMS and to check up about presence in it of it Pack Key.
2).  groups of measures, there by errors - the request similarly derived, looked at it. If it is necessary - launched in SSMS, found the specified line and checked up that there not so.
How all it to make?

5

Re: Error at handling of cube Analysis services 2008R2 standard

WarAnt wrote:

Ferdipux;
Daw materialize at these measurements remove and there will be to you a happiness

Thanks, Ferdipux. Where there is this daw?

6

Re: Error at handling of cube Analysis services 2008R2 standard

ondorsal wrote:

it is passed...
Thanks, Ferdipux. Where there is this daw?

Daw found, but I do not have referential communications in measurements, the problem remains.

7

Re: Error at handling of cube Analysis services 2008R2 standard

ondorsal wrote:

the Daw found, but I do not have referential communications in measurements, the problem remains.

Separately  measurement "preparations" (update type)
Pluses , in them look a SQL query
It fulfill in studio, with the necessary key

8

Re: Error at handling of cube Analysis services 2008R2 standard

StarikNavy wrote:

it is passed...
Separately  measurement "preparations" (update type)
Pluses , in them look a SQL query
It fulfill in studio, with the necessary key

Checked up, launched request of handling of measurement "Preparations" and really there there is no line c necessary ID.
In request it is a lot of join tables, but all of them inner-see they cut the lines necessary to me. How in at processing to change type  on left join?

9

Re: Error at handling of cube Analysis services 2008R2 standard

ondorsal wrote:

As in at processing to change type  on left join?

In any way. It is meant that referential integrity is supervised in ETL or somehow at level of the sources used in datasource view (but it not is  ).

10

Re: Error at handling of cube Analysis services 2008R2 standard

Theoretically, it is possible to disconnect error control of absence of a key in measurement at cube processing, but it already absolutely  .

11

Re: Error at handling of cube Analysis services 2008R2 standard

alexdr wrote:

Theoretically, it is possible to disconnect error control of absence of a key in measurement at cube processing, but it already absolutely  .

Thanks. Altered measurement, constructed on c view with necessary . Now all works.

12

Re: Error at handling of cube Analysis services 2008R2 standard

alexdr wrote:

it is passed...
In any way. It is meant that referential integrity is supervised in ETL or somehow at level of the sources used in datasource view (but it not is  ).

datasource view (but it not is  ) Why badly to use representations?

13

Re: Error at handling of cube Analysis services 2008R2 standard

ondorsal;
It is considered that before "consumption" the data should be well under (at) . And an interlayer between a cube and DWH in a type I twist, which do not do additional operation - why are not present?

14

Re: Error at handling of cube Analysis services 2008R2 standard

alexdr wrote:

Theoretically, it is possible to disconnect error control of absence of a key in measurement at cube processing, but it already absolutely  .

Normal practice. If it is admitted that reference manuals  later.
And then it is not necessary superfluous  in , for check of keys

15

Re: Error at handling of cube Analysis services 2008R2 standard

WarAnt wrote:

Ferdipux;
Daw materialize at these measurements remove and there will be to you a happiness

It not referenced dimension, as at . It when in measurement some tables from DSV with the communications defined in between are pulled out. SSAS constructs requests, and the logic of a choice join - left or inner - is clear not always. Somehow did attempt to understand, like it is connected to property CanBeNull of a field in the parent table, but it anywhere explicitly is not described.

16

Re: Error at handling of cube Analysis services 2008R2 standard

StarikNavy wrote:

normal practice. If it is admitted that reference manuals  later.

I hold other judgement. Does not mean better easier. Often it is necessary to pay off for it expensively. And for handling of a situation with Late Arriving Dimensions (where without them?) there are for a long time known and well described patterns. They also are - best practice.

17

Re: Error at handling of cube Analysis services 2008R2 standard

alexdr wrote:

often it is necessary to pay off For it expensively.

If it is not difficult, it is possible an example of expensive payment?
(Without sarcasm, it is really interesting)

18

Re: Error at handling of cube Analysis services 2008R2 standard

As a variant, the first that got:
[quote =] If you want to process the database or the cube without correcting the data, you can set the error configuration for the process operation to ignore the error. You should only do this as a temporary workaround when you fix the underlying data. Otherwise, you may receive unexpected results from your multidimensional expressions (MDX) queries.

Unexpected results conduct to that users cease to trust the data from OLAP, as a result all project in which so much forces are enclosed, abilities and so forth easily can be  since it is not claimed by users because there - lie.
The second. Absolutely not mandatory that a unique customer of the data is MS OLAP cubic Hence, such "dirty " should be applied in all customers of the data what often to observe very difficult, especially when for development and BI-decision support answers more than one (generation) of experts.
Somehow so...