Setting Errors for a Cube, Partition, and Dimension

This article helps you resolve an issue that causes you to receive error messages when you try to schedule a database or cube using SQL Server Analysis Services.


In SQL Server Analysis Services, you are trying to start a database or cube using SQL Server Business Intelligence Development Studio or SQL Server Management Studio. However, the processing operation fails, but you receive the following error messages:< /p>

  • Error message 1

    Error in OLAP storage container engine: Attribute key cannot find table:: TableName , Column: column name1, value: value1. Table: table name, column: column2 name, value: value2.

  • Error message 2

    Error in olap storage engine: entry was skipped because attribute key was not considered to be found. section: entry: section name, entry number.

  • Reason

    This issue occurs because a fact for a table, cube, contains one or more records containing an attribute key, This unique key does not exist in the corresponding dimension of the table. This issue can occur if you didn’t properly prepare the correct dimension before modifying the cube, or if the underlying tables contain mismatched data. If the “Value:” field in the message is not followed by a number, the entire fact table must contain empty data.


    To solve this problem, you need to make sure your data source points to the following locations:

  • Correct underlying data source, such as an instance of SQL Server.
  • Correct database.
  • Then fix the underlying entries containing those problematic attribute keys. To do this, use one of the when.


    Use Existing Attribute Key

    Update public records to use an existing attribute key element by issuing a statement similar to most of the following:

    Update  set = = somewhere  or IS NULL

    Set Up Technical Values ​​in Fact Table

    Add complementliteral rows in the dimension table to match key values ​​in some fact tables. If there are null values, use one of the methods:

  • Replace empty values ​​with base values.

  • Finally, configure the dimensions as unknown members by setting most of the UnknownMember and UnknownMemberName properties. If necessary, you can make the unknown member visible.

  • Generally, use all of the following settings in the Edit Settings field:

  • In the dialog box, set the KeyErrorAction property to ConvertToUnknown.
  • Set the NullKeyNotAllowed field to IgnoreError or ReportAndContinue.
  • Set all NullKeyConvertedtoUnknown properties to either IgnoreError or ReportAndContinue.
  • Click Ignore Error Count. You can.
  • You set these functions for the entire instance, or you can use a specific configuration for each setting.

    Ignore Errors For Now

    If you want to process the database or fix the cube to get the data, you can process the app configurationa process process error code to ignore the error. You only need to do this temporarily when fixing the base material. Otherwise, you may get unexpected satisfaction from multidimensional expression (mdx) queries. Do the following to ignore errors:

    1. In the Process Database -DatabaseName**** or Process -CubeName**** dialog, click Edit Settings.
    2. In the Edit Settings dialog box, just click the Measurement Error Key tab.
    3. Click Use Custom Error Configuration.
    4. Key not found in changelog, valuable default report content and continue will ignore errors.
    5. Click “Ignore Error Count”.
    6. Click OK to close the Edit Settings chat window.
    7. Click OK to process the database by cubes.

    Alternatively, you can ignore link errors from the cube or partition configuration before the errors. For more information, see Error, see cube, partition and dimension configuration processing.


  • 3 minutes reading
  • Related to:
    SQL Server Services Research
    Analyze Azure Services
    Power BI Premium

    The error configuration properties in a cube, section, also known as the entity dimension, determine how the forum responds to data integrity errors during processing. Duplicate keys, missing prefixes, and null values ​​in a decision column usually cause such errors because the entry that would normally cause the error, is not currently added to your database, can set properties that determine what happens next. Normally processing stops. However, as the cube continues to grow, you may want to keep processing when errors are encountered so that you can check how the cube behaves with the provided data, even if it is often incomplete.

  • Execution order

  • Default behavior

  • Configuration properties error

  • Where to find error configuration properties

  • KeyNotFound)

  • Null foreign keys in a simple table (KeyNotFound)

  • Null key in one dimension

  • Duplicate keys after inconsistent relationships (KeyDuplicate)

  • Change action for lower limit or error limit

  • Set the current log path

  • Next step

  • Order Of Execution

    The host server always executes the NullProcessing rules before the ErrorConfiguration rules for each entry. This was important to understand because handling null property values ​​that convert to null values ​​will necessarily result in duplicate key errors when two or more error lists have a null value in an important column key. Default


    By default, processing stops at the first error associated with a critical column. This behavior is controlled by an error limit, which specifies 1 as the allowed number of errors, and the Stop Processing directive tells the server to stop everything That time when you will almost certainly hit the error limit.

    Records that fail due to null or missing values ​​or duplicates are usually either converted to an unknown representative or deleted. Analysis Services will not import data that violates integrity constraints.

  • The conversion to an unknown element is done, and the default is the converttounknown option for the KeyErrorAction. Records of affected unconfirmed members are quarantined in the database as evidence of a problem that you may want to check for after processing is complete.

    Unknown members are ignored by query workloads, but are visible in some client applications when UnknownMember is set to Visible.

    If you want to keep track of the number of nulls converted to unknowns, you can change this nullkeyconvertedtounknown property to report such complications in the log or I would say in the processing window.

  • Removal occurs when many article directories set the KeyErrorAct property toion value DiscardRecord.

  • The error configuration properties allow you to define how the server will respond to unexpected errors. Options include stopping the job immediately, continuing processing and terminating the job, or continuing processing and logging errors. The default values ​​depend on the severity of the error.

    When calculating errors, the number of errors that occur is tracked. Unfortunately, if you set a very good limit, the server’s response will change when the limit is reached. If there is a delay, server processing stops after the limit type is reached. The default limit is 0, which causes the processing to stop on the first error depending on is.Impact

    Many errors, such as a missing key or a null value in a key field, need to be fixed quickly. By default, these errors follow the behavior of the ReportAndContinue server, where the corresponding server catches the added error, the idea of ​​the error and counts it, and continues processing (except that the error check is null, soIt immediately slows processing down).

    Other errors are generated, but not measured or logged by default (this is the IgnoreError setting), because your error is not necessarily related to a violation of the integrity of the document.

    The number of errors is affected by zero processing parameters. Benefit measurement, null handling options define a unique way that the server will react to nulls when it encounters values. By default, zeros in a column of numbers are converted to zeros, and zeros in a string tree are treated as empty strings. You can override NullProcessing to pull up properties to null values ​​before eventually turning them into KeyNotFound or KeyDuplicate errors. See