When staging tables are dropped and re-created, the best version of the truth (BVT) calculation does not consider the Allow Null Update property settings for the staging table columns. When a staging table is recreated, the cross-reference records loaded by the original staging table during BVT calculation are affected.

If a foreign key value is updated with a null value when the option to update cells is enabled in the staging table, the load job fails. The reject table is not populated with rejected records and an error message does not appear.

In a Microsoft SQL Server and IBM WebSphere environment, if you enable distinct mapping load jobs fail with the following error:

DataAccessException: SIP-10322: Unable to execute the SQLInvalid column name ROWNUM

In a Microsoft SQL Server environment, load jobs fail if records contain empty column values.

Match jobs with large batch sizes take a long time to run.

In the Microsoft SQL Server environment, if you use a database port number other than the default, which is 1433, Operational Reference Store (ORS) upgrades fail.

When you unmerge records, the merge update event trigger entry in the C_REPOS_MQ_DATA_CHANGE table contains incorrect source systems in the SRC_ROWID_SYSTEM and TGT_ROWID_SYSTEM columns.

If you use a Korean population, and run a match job for Korean data, the Process Server fails.

When a merge job runs, the MDM Hub generates incorrect best version of the truth (BVT) for records that undergo consolidation.

When you soft delete a cross-reference record, the base object is soft deleted even if other cross-reference records are active.

When using IBM DB2 for the Hub Store, messages are not sent to the JMS queue.

A composite service request populates the create date column of a base object with the current date instead of the create date that is passed in the composite service request.

The PromotePendingXrefs API might not promote the correct cross-reference records, which results in base objects with incorrect data updates.

When you use the Automerge job to delete the match records for all base object records with consolidation_ind=9, the performance is unacceptably slow. The issue occurs when the database contains a million affected base object records and the match table contains greater than 2 million records.

When you run a load job, the MDM Hub rejects records for timeline-enabled base objects and generates the following error:

When the Allow non-contiguous effective periods option is disabled for a base object and you try to add cross-references with different effective periods to a record, the MDM Hub generates the following error message:

If you run a batch unmerge job on a child record with more than 1000 cross-reference records, the MDM Hub generates an exception.

If you make a Put API call to update a child record, the foreign key of the parent base object might point to the wrong cross-reference record.

If the effective period of the parent record precedes the effective period of the subject area child, then task creation fails with no error message.

When you update an entity in the Hierarchy Manager, the following error occurs:

In the Microsoft SQL Server environment, the MDM Hub incorrectly detects the records that are hard-deleted in source systems.

When you enable the Match Subtype property on a match column, some unrelated records might be matched.

After upgrading from Informatica MDM Multidomain Edition version 9.5.1 HF1, support for Oracle VPD is missing.

When the Cleanse Reject function rejects a stage record, the reject table does not include the SRC_ROWID value of the rejected record.

You can see the New drop-down menu in Informatica Data Director without assigning create privileges on the subject area of the associated base object.

When a stage job loads unicode data that is rejected for exceeding the column length, the reject table is not populated with the rejected records

If you do not select a package when you configure a message trigger, the message queue displays the following error message:

When you delete a record that has cross-references from different sources, some fields in the cross-references records can change.

When you run a Stage job with the Enable distinct option enabled in mapping, some records might not be inserted in the staging table.

When you run a match job in parallel with real-time updates, the match job fails.

When you run a load job on a base object where the stage table is cell update enabled and has a lookup to another base object, the load process generates the following error:

When you enable Delta Detection and Allow staging if prior duplicate was rejected on a staging table, the stage job does not stage the records from the reject tables.

The database upgrade script adds the prefix CLM% to the display name of system columns and user-defined columns.

You can run a Put call on a child base object at the same time that you run a match job on the parent base object. If there is a match path from the parent base object to the child base object, a locking issue occurs on the parent DRTY table.

If you update a base object with multiple cross-reference records, the MDM Hub sends multiple messages to the message queue.

If you enable cross-reference columns for case insensitive search, and then create more cross-reference columns, the MDM Hub nullifies the newer cross-reference columns when you run an automerge job.

A match analyze job might fail with the following error:

When you call executeBatchUnmerge SIF from a SOAP interface, errors are generated in the log file.

When you save a custom query, the query might become unresponsive.

When running

sip_ant updatemasterdatabase

, the database migration script fails with the following error:

ORA-00959: tablespace 'CMX_IND' does not exist

When you mark columns as putable in the base object table, the load process can incorrectly write null values to the putable columns in a base object record and a cross-reference record.

The match batch process fails when real-time operations run in parallel.

When you have multiple Operational Reference Stores (ORS) with the same name and you import the user exit JAR file into each ORS, the Process Server fails to launch with a fatal exception.

When you run an unmerge batch job, multiple cross-references associated with the same base object cannot unmerge and an error occurs.

After an edit to a record is approved, when you edit the same record again, a NULL value is written to the interaction_id column of the C_REPOS_TASK_ASSIGNMENT_HIST table. The issue occurs when you use Siperian BPM and the Hub Store is in Microsoft SQL Server or IBM DB2.

When you have multiple Operational Reference Stores (ORS) with the same name and you import the user exit JAR file into each ORS, the Process Server fails to launch with a fatal exception.

When you update or promote a package from the Hub Console, all user privileges that you granted on the corresponding Oracle view are removed.

A match analyze job might fail with the following error message:

Comparison method violates its general contract! 
				

For a base object record, you can set up a message trigger for multiple source systems. For each update to the base object record, multiple messages are published to the queue instead of a single message.

You can assign batch group permissions to a user who is not an administrator. When this user logs in to the Hub Console from a remote computer, it can take up to 15 minutes to display the batch groups.

A failed job in a batch group does not get logged in the job control table.

The metadata validation process generates the following error:

When you specify the limitDate parameter for the ExecuteBatchExtractBVTVersions element in a request, the timeline extract process ignores the time portion of the limitDate parameter.

After you enable a database user exit, the load process fails with the following error:

When you mark columns as putable in the base object table, the load process can incorrectly write null values to the putable columns in a base object record and a cross-reference record.

After you run a timeline extract process, the extracted records have blanks or NULLs in the creator, create_date, and updated_by_fields columns.

Custom queries that contain greater than 4,000 characters cause the searchQuery API to fail with a syntax error.

When you run a batch group that runs multiple stage jobs in parallel against the same landing table, the stage jobs fails with the following error:

You can assign batch group permissions to a user who is not an administrator. When this user logs in to the Hub Console from a remote computer, it can take up to 15 minutes to display the batch groups.

For each insert or update, more than one message is published on the message queue.

An unmerge event might not trigger any outbound message.

The cmxserver and cleanse logs are not rolling over when they reach maximum size. As part of this fix, new log files were added. For more information about logging, see the

Configuration Guide

.

After the upgrade, the generic Service Integration Framework (SIF) WSDL that is generated is not valid.

When you run an unmerge batch job, multiple cross-references associated with the same base object cannot unmerge and an error occurs.

When an update task is created and approved, the task assignment history table, c_repos_task_assignment_hist, is not updated with a valid Interaction ID value.

If multiple columns contribute to the primary key, and if one of the columns from which the primary key is derived is null, stage job fails if delta detection is enabled.

You cannot preserve source system keys during load jobs that are run after the initial load job.

Some SearchMatch requests fail due to a synchronization issue with SearchMatch worker threads. The issue occurs randomly in a fast-paced environment.

If data from multiple source systems exist in a timeline-enables base object, you cannot get the best version of the truth (BVT) of past, present, and future data.

When you open the batch viewer and then reduce the size of the Hub Console, some information is off screen and there is no scroll bar.

After merging records, preexisting tasks do not become obsolete. The tasks continue to appear in the Task Inbox in Informatica Data Director. If you open one of these tasks, you see an error.

During the stage process, the performance of the delta detection operation is poor.

In the History view, the wrong event details are displayed for lookups when the base object record row ID is different than the cross-reference record row ID.

After you load data that has a higher trust value than the existing data, the trust calculation is incorrect.

When changes to effective dates are not contiguous with the original effective dates, the Mark for Save action fails.

If you use the Promote API but do not provide a source key, the MDM Hub runs an SQL statement that results in a full table scan. Consequently, performance becomes slower.

When you perform Informatica platform staging and there is more than one Operational Reference Store registered in the Hub Console, metadata might be inaccurate. Updates to staging table metadata are inaccurate if there are base objects with the same name in different Operational Reference Stores.