SR-D93777 · Issue 565692
Handling added for Oracle Aggregate IH Summaries
Resolved in Pega Version 8.2.8
When using using (Non Materialized) IH Summaries to aggregate IH data, the data returned by the IH summary did not include all the expected records. If the same criteria was executed on the database via SQL or by using a strategy to process raw IH data then the results were as expected. This was due to a difference in handling of Oracle vs postgres which causes an order by clause not to be generated in the query: the postgres column name is lower case, while in Oracle it is upper case. This has been resolved by updating the system to get the column name correctly from the propertytocolumn map so IH records are returned in correct order by Browse By Keys operation.
SR-D95605 · Issue 565486
Data Flow correctly saved to Database table Dataset
Resolved in Pega Version 8.2.8
After setting up a dataflow with Report definition as a source and Database table dataset as destination with the option "Insert new records and override existing", a data transform was used to modify a few of the values and write to the same database table. This table was mapped to pegadata schema and didn't have a pzpvstream column. Running the data transform generated an exception stating "DataStoreSaveStatementWithoutStream(PageDatabaseMapperImpl). The error was not seen when the pzpvstream column was added or if pzInskey was also removed along with pzpvstream, or if "only insert new records" was selected. This was traced to pzInsKey and pxInsName being null in the query formed while writing. There are two execution paths for the database dataset save operation, one for internal tables and another one for external tables. For a table with pzInskey but no blob, the system was incorrectly using the external table logic. This has been corrected.
INC-121460 · Issue 569118
IH Summaries correctly loads aggregated count
Resolved in Pega Version 8.4.2
After the IH Aggregated DataFlow has processed more than 34 million records from an Oracle database, the IH Fact table contained only ~120,000 records. Running a browse on the pxInteractionHistory DS resulted in an infinite loop. Investigation showed that this was caused by the handling of Oracle timestamps: using Pega DateTime in query parameters generated Oracle DATE instead of TIMESTAMP, causing all query parameters to be rounded to the second and interfering with the IH pre-aggregation. This has been resolved by updating the handling.
INC-122180 · Issue 563270
Null check added for Entity Extraction result
Resolved in Pega Version 8.4.2
Training data for an Entity Extraction model was not getting distributed correctly, as per the given % data. This was due to the result of the Entity Extraction result being null, and has been resolved by adding a check that will consider detected types as primary candidate for grouping if the result column is null.
INC-125838 · Issue 561484
Ignore Case feature updated
Resolved in Pega Version 8.4.2
In Pega Marketing 8.4 on the Action rule form, configuring an eligibility criteria and selecting the option for the comparison to ignore case did not copy this part of the criteria to the back proposition filter rule, so the case of the comparison was not ignored at runtime. This has been corrected.
INC-125900 · Issue 560788
ResolveFTCR flow Draft mode turned off
Resolved in Pega Version 8.4.2
When using the FastTrack Change Request functionality of Revision Manager 8.3 the submission was going to Pending packaging instead of Resolved-Completed with the error "flow is in draft mode; it cannot be executed". This was due to pxResolveFTCR flow having been checked in with Draft mode turned on, and has been corrected by turning Draft mode off.
INC-126129 · Issue 569668
PropertyToColumnMap made more robust
Resolved in Pega Version 8.4.2
The DF_ProcessEmails dataflow was intermittently failing with a StageException error. This was traced to schema changes being propagated asynchronously by system pulse, which seem to have caused PropertyToColumnMap to cache stale schema. To resolve this, if the property mapping is not found the first time, the system will make a second attempt to get the mapping. Additional logging has also been added for better diagnostics.
INC-126187 · Issue 564340
Report Definition replaced with a data set for training the data and build model
Resolved in Pega Version 8.4.2
An issue with training the data and build model via an email or chat channel was traced in an exception in logs indicating the inability to execute a Report Definition. This can happen when UDF are removed/disabled in the database or in environments where functions are not allowed to read the blob for performance reasons. To resolve this, the Report Definition has been replaced with a data set.
INC-126750 · Issue 564155
DSS added to configure Dataset-Execute page handling
Resolved in Pega Version 8.4.2
When Kafka Data Set pages were saved to a data set with the Dataset-Execute method, there was is feedback if any of the pages were not successfully saved. Instead, the step always completes as successful. In addition, if any properties are added or modified by the save operation itself, those changes are not visible. This is due to the data set execute save operation saving pages as DSM pages to the data set. Due to the conversion of the pages, copies of the pages are used which do not reflect back any changes on the input pages. DSM pages are used by default because they are more lightweight than regular clipboard pages and therefore have potentially better performance. In order to allow the use of DSM pages to be customized, a new Dynamic System Setting: dataset/execute/save/statusFailOnError has been added. This can be enabled by setting it to true; it is disabled by default for greater backwards compatibility. By removing the DSM page conversion in the generated save code, changes to input pages will be reflected back if any are performed by the data set save operation, and the system will report back which pages are saved or failed by adding messages to the pages that failed to save. Performance may be affected with this change as regular clipboard pages are in general slower than DSM pages, however, that may be offset by removing the conversion to DSM pages process and will depend on the site configuration.
INC-126796 · Issue 561536
Modifications to getFunctionalServiceNodes process
Resolved in Pega Version 8.4.2
The count of the Interaction History write related threads was increasing rapidly and a stack trace indicated "waiting on condition" and "java.lang.Thread.State: WAITING (parking)" errors. Investigation showed that this was due to getFunctionalServiceNodes using Hazelcast to determine node status by making a service request on an installation with a very large number of nodes, causing thread locking. To resolve this, the implementation has been updated to avoid calling getFunctionalServiceNodes on save of Interaction History, instead using Cassandra and only calling getFunctionalServiceNodes on the master node, not on all nodes.