What are the differences between the informatica MDM features of Oracle, SqlServer and DB2?

In master data management (MDM), there is a difference in how databases work on the database side.

In Oracle, an empty string (") is treated as a NULL value, but in SqlServer and DB2 they are two separate ICM attributes. An empty string ("") is treated as "Null " is converted to

For SqlServer and DB2, MDM accepted empty strings by converting them to null values.

For example, the MDM package (database view) uses the NULLIF database function to return null for an empty string. The data types in SqlServer and Oracle are different, but there are equivalents. For example:

Oracle number = SqlServer number. Character data types in SqlServer are NVARCHAR, NCHAR, but Oracle uses character length VARCHAR2 and CHAR. The date columns have different data types.

The SqlServer date type used in MDM is DATETIME2(7) or DATETIME2(0). Oracle uses TIMESTAMP and DATE respectively, but there are equivalents. DB2 datatypes are very similar to Oracle datatypes.

What are your recommendations for improving Elastic Search in Informatica MDM?

Elastic Search parameters can be tuned to improve performance.

We recommend that you use the same value for the Xmx (maximum memory) and Xms (minimum memory) Java parameters in your Elastic Search configuration.

If the loading job takes a long time to intelligently search indexing for a particular model, set this property to better optimize indexing. cmx.ss.index.build.optimization=true . For more information, see How to: Improve load job performance when smart scan is enabled in MDM (KB 000156763).

Based on your application server performance, increasing this value will improve index batch job performance in the cmxserver.properties file (cmx.server.batch.smartsearch.initial.block_size). The default value is 250. Recommended value is 5000.

Maximum number of threads distributed across different process servers to process batch jobs and initially index smart scan data. Increasing this value will improve the performance of this batch job. This property does not apply to regular indexing outside of this particular batch job. Same properties reused from the LOAD job. Set this in the cmxserver.properties file cmx.server.batch.threads_per_job.

Communication between the app server and Elasticsearch can be lost when indexing jobs run on a large number of records. Add the JVM property -Dmdm.es.index.client.ttl=864000 when starting the server. This property specifies the number of seconds the application server should wait before closing old connections on the Elasticsearch server. Informatica recommends using 86400 seconds. The default value is 7200 seconds which is approx. 2 hours.