Known limitations and workarounds in InfoSphere Information Analyzer Version 11.5 fix packs and rollup patches

Known issues in InfoSphere Information Analyzer, Version 11.5 Fix Pack 2 Service Pack 1:

1. Data rule jobs cannot be run if they include columns with INT8 and INT16 data types and the source repository is SQL Server that is being accessed by using the ODBC Connector and the Information Analyzer database (IADB) was setup with DB2 and is accessed by using DB2 Connector. To work around this issue, configure the Information Analyzer database (IADB) so it can be accessed by using the ODBC Connector.

2. You try to save the ERD diagram to a PDF document but the document is blank or does not print properly. To work around this issue, verify that you have a web browser plugin, such as CutePDF or Adobe Acrobat in order to save the ERD diagram to a PDF format. Also, check the settings for the PDF writer plugin.

3. You cannot test or validate data rules in the thin client that have expressions in the "Define output" tab that contain date related functions. To work around this issue, delete the date related functions from the "Define output" tab of your expression.

4. A virtual column cannot be created within a SQL virtual table. To work around this issue, you can build the the equivalent of a virtual column within the query used to create a SQL virtual table.

Known issues in InfoSphere Information Analyzer, Version 11.5 Fix Pack 2:

1. Browser limits on graphical display of relationships - Due to limits on various web browsers the total number of data sets graphically displayed at one time has been limited. There is also a warning if the total number of relationships displayed exceeds a limit. Both limits vary by web browser type and both are configurable by an Administrator. Note: Users can easily select/deselect data sets so they may focus on one or a few at a time. This typically avoids hitting these limits. In the cases where the relationships can not be rendered graphically, there is always a grid/list view that shows all relationships along with the details of the strength and confidence of each.

The default limits are:

- Firefox - 30 data sets (graph limit), 50 relationships (warning)
- Internet Explorer - 10 data sets (graph limit), 20 relationships (warning)

The Administrator can influence these settings as described in the IBM Knowledge Center topic "Setting properties from a command line" via the following properties:

com.ibm.iis.ia.client.cda.diagram.dataSetLimit (Firefox) - hard limit on ERD entity count - default=40
com.ibm.iis.ia.client.cda.diagram.relationshipLimit (Firefox) -soft limit, confirmation/warning - default=60
com.ibm.iis.ia.client.cda.diagram.dataSetLimit.ie (IE) - hard limit on ERD entity count - default=25
com.ibm.iis.ia.client.cda.diagram.relationshipLimit.ie (IE) -soft limit, confirmation/warning - default=45

Please be sure to refer to the System Requirements for the list of supported Web Browsers and versions http://www-969.ibm.com/software/reports/compatibility/clarity/softwareReqsForProduct.html

2. Slow response times can be expected in the Information Analyzer thin client for Relationship Analysis when more data sets and relationships are involved because of high CPU usage on the Engine tier.

3. The algorithms used in the background during Key & Cross Domain analysis in Information Analyzer Workbench and the Relationship Analysis in Information Analyzer thin client are different. You may experience inconsistency when you try to run the analysis in one client and view the results in the other. You should run and view the results in the same client.

4. Users will not be able to drill down / view sample records from a Hive repository when the column data type is "timestamp."

5. The output definition grid in the Information Analyzer thin client does not show variable bindings correctly when a data rule is edited.

6. When a variable of a Datarule is bound to a Literal which is created using a Date value, the value of literal displayed is 1 day less than that of the value entered.

ex, User enters '2017-01-01' but the value displayed is '2016-12-31'

7. Column Analysis on a file (containing less than 10 columns/fields) fails when one or more Virtual Column(s) defined and analysis is performed in one go using a single job. Alternatively users can run column analysis on fewer columns in multiple runs from workbench.

8. File records that contains double quotation marks ("") are skipped during a data quality analysis and data rule executions.

9. From InfoSphere Metadata Asset Manager (IMAM), if you delete a dataset that is previously analyzed in the thin client, you will not be able to view that dataset in the Workspace or Find Data screen of the thin client. If you need to view the analysis results of the dataset that is removed, use the IA workbench.

10. It is possible to save a data connection that has been created using incorrect information, such as incorrect host name, user name, or password. To avoid data connection errors, test and validate all data connections before saving.

11. Users assigned only the user role "Common Metadata User" cannot access file based connections. Add an assignment of "Common Metadata Administrator" to allow access to these connections.

12. Data set or column descriptions edited in Information Analyzer workbench will not be reflected in the thin client. To see modifications in both the workbench and the thin client, make changes in InfoSphere Information Governance Catalog.

13. When the "city classifier" data class is enabled, you can expect performance degradations during data quality analysis. If you are concerned with performance and do not have a need for the "city classifier", disable the data class using the workbench or Information Governance Catalog.

14. An analysis job might fail with the message below if you run an analysis that includes columns that have a string data type with an unspecified length:

Symptom:
=======
You see the following message in logs after an analysis:
"DataStageSession.java:162 Event 1: main_program: Fatal Error: Size arg must be less than 1,048,576"

Workaround:
==========
Define the additional parameter 'MaxStringSize' in JDBC URL with appropriate size in the DataConnection and then import metadata from source.

example: jdbc:ibm:hive://<HIVE-SERVER>:<PORT>;MaxStringSize=256

PS: The value for MaxStringSize can be extended depending on the maximum length of the string values exists in the source Hive system.

15. Users are unable to perform Column/ Data Quality analysis for high volume source data when the Information Analyzer analysis database (IADB) is connected by using the DB2 connector or the ODBC connector.

Symptom:
=======
pxbridge(4),5: [Input link 0] SQLConnect reported: SQLSTATE = 57011: Native Error Code = -956: Msg = [IBM][CLI Driver] SQL0956C Not enough storage is available in the database heap to process the statement. SQLSTATE=57011 (CC_DB2Connection::connect, file CC_DB2Connection.cpp, line 852)

pxbridge(3),1: ODBC function "SQLConnect" reported: SQLSTATE = HY000: Native Error Code = -956: Msg = [IBM(DataDirect OEM)][ODBC DB2 Wire Protocol driver][DB2]Unknown error: SQLCODE -956 (CC_OdbcConnection::connect, file CC_OdbcConnection.cpp, line 584)

Workaround:
==========
Modify the DBHEAP parameters using the following command after connecting to Information Analyzer Analysis Database (IADB):

UPDATE DB CFG FOR IADB USING DBHEAP AUTOMATIC

16. Column Analysis and Data Quality Analysis fails with the following specific error when Metadata is persisted in SQLServer repository and domain tier is Websphere Liberty.

Symptom:
=========
messages.log contains the following error for the specific execution:

<string>pxbridge(1),0: java.util.regex.PatternSyntaxException: Dangling meta character &apos;?&apos; near index 105
^lat$|^lat_|_lat$|latitud(ine|e|o)?|breitengrad|breddekreds|breedtegraad|breiddegrad|breiddargráða|enlem|??????|?????(?|a)|?? ????|??? ??????|??|??

Workaround:
===========
Open Governance Catalog -> Information Assets -> Data Classes and disable Latitude and Longitude and re-run the column analysis or data quality analysis.

Known issues in InfoSphere Information Analyzer with Governance Rollup 7:

1. Users will not be able to create HDFS data connection in Information Analyzer thin client when more than one MIS server is enabled in Metadata Asset Manager. The work around to get over the issue is i) to disable client only MIS and work with Information analyzer thin client ii) delete the Host (client MIS hostname) from repository

2. Browser limits on graphical display of relationships - Due to limits on various web browsers the total number of data sets graphically displayed at one time has been limited. There is also a warning if the total number of relationships displayed exceeds a limit. Both limits vary by web browser type and both are configurable by an Administrator. Note: Users can easily select/deselect data sets so they may focus on one or a few at a time. This typically avoids hitting these limits. In the cases where the relationships can not be rendered graphically, there is always a grid/list view that shows all relationships along with the details of the strength and confidence of each.

The default limits are:

- Firefox - 30 data sets (graph limit), 50 relationships (warning)
- Internet Explorer - 10 data sets (graph limit), 20 relationships (warning)

The Administrator can influence these settings as described in the IBM Knowledge Center topic "Setting properties from a command line" via the following properties:

com.ibm.iis.ia.client.cda.diagram.dataSetLimit (Firefox) - hard limit on ERD entity count - default=30
com.ibm.iis.ia.client.cda.diagram.relationshipLimit (Firefox) -soft limit, confirmation/warning - default=50
com.ibm.iis.ia.client.cda.diagram.dataSetLimit.ie (IE) - hard limit on ERD entity count - default=10
com.ibm.iis.ia.client.cda.diagram.relationshipLimit.ie (IE) -soft limit, confirmation/warning - default=20

Please be sure to refer to the System Requirements for the list of supported Web Browsers and versions http://www-969.ibm.com/software/reports/compatibility/clarity/softwareReqsForProduct.html

3. Slow response times can be expected in Information Analyzer thin client for Relationship Analysis when more datasets and relationships are involved because of high CPU usage on the Engine tier.

4. The algorithms used in the background during Key & Cross Domain analysis in Information Analyzer Workbench and the Relationship Analysis in Information Analyzer thin client are different. You may experience inconsistency when trying to run the analysis in one client and view the results in the other. It is suggested to Run and View the results in the same client.

5. The project XML generated using getProject command of IAAdmin script for projects registered with file based data sources contain the same data connection info even when different files are imported using different data connections from same host. Since all these connections refer to the same details there is no impact of this during analysis jobs

6. File records that contains double quotation marks ("") are skipped during a data quality analysis and data rule executions.

7. From InfoSphere Metadata Asset Manager (IMAM), if you delete a dataset that is previously analyzed in the thin client, you will not be able to view that dataset in the Workspace or Find Data screen of the thin client. If you need to view the analysis results of the dataset that is removed, use the IA workbench.

8. It is possible to save a data connection that has been created using incorrect information, such as incorrect host name, user name, or password. To avoid data connection errors, test and validate all data connections before saving.

9. Users assigned only the user role "Common Metadata User" cannot access file based connections. Add an assignment of "Common Metadata Administrator" to allow access to these connections.

10. Data set or column descriptions edited in Information Analyzer workbench will not be reflected in the thin client. To see modifications in both the workbench and the thin client, make changes in InfoSphere Information Governance Catalog.

11. When the "city classifier" data class is enabled, you can expect performance degradations during data quality analysis. If you are concerned with performance and do not have a need for the "city classifier", disable the data class using the workbench or Information Governance Catalog

Known issues in InfoSphere Information Analyzer thin client with Governance Rollup 6:

1. When the user has more than 25 rules in a workspace, the user may not be able to see all rules in certain versions of Firefox.

2. Drill down from "Data rule status" or "Data rules success rate" charts in workspace summary doesn't show any results, or in some cases shows incorrect results. Workaround is to rebuild data set index. Refer to http://www.ibm.com/support/knowledgecenter/SSZJPZ_11.5.0/com.ibm.swg.im.iis.ia.product.doc/topics/t_reindex_daui.html for details.

3. From InfoSphere Metadata Asset Manager (IMAM), if you delete a dataset that is previously analyzed in the thin client, you will not be able to view that dataset in the Workspace or Find Data screen of the thin client. If you need to view the analysis results of the dataset that is removed, use the IA workbench.

4. Performance on Oracle IADB/Xmeta could be slow and might cause rule grids to take several seconds to load.

5. It is possible to save a data connection that has been created using incorrect information, such as incorrect host name, user name, or password. To avoid data connection errors, test and validate all data connections before saving.

6. File records that contains double quotation marks ("") are skipped during a data quality analysis and data rule executions.

7. Users assigned only the user role "Common Metadata User" cannot access file based connections. Add an assignment of "Common Metadata Administrator" to allow access to these connections.

8. Data set or column descriptions edited in Information Analyzer workbench will not be reflected in the thin client. To see modifications in both the workbench and the thin client, make changes in InfoSphere Information Governance Catalog.

9. When the "city classifier" data class is enabled, you can expect performance degradation during data quality analysis. If you are concerned with performance and do not have a need for the "city classifier", disable the data class using the workbench or Information Governance Catalog.