Reltio Cloud 2019.3 Weekly Releases

This page lists bug fixes and enhancements applied as part of our Weekly Releases in Reltio Cloud 2019.3.

Release Notes for 2019.3.0.17 Emergency Bug Fix Release - Feb 10, 2020

Fixes

Hub

Users using Microsoft Edge on a device with a touch screen could not select or de-select the bar charts on the search facets. This issue is now fixed.

Salesforce Connector

  • Contacts with more than 130 nested appointments were not synchronized even after manual update. This issue is fixed by using the Salesforce Bulk API. Now, synchronization of big entities is not failing.
  • Nested attributes merged due to change in entities were only merged in Reltio and not in Salesforce. As a result, merge inconsistencies were observed between Reltio and Salesforce. This issue is now fixed by modifying the data calculator so that it can identify the merged entities in Salesforce as well.
  • Merges of all types were processed by the Salesforce Connector even if the mergeTypes property was empty in the mapping section. As a result, the mergeTypes field was ignored while mapping. This issue is now resolved.

Release Notes for 2019.3.0.16 Weekly Release - Jan 30, 2020

What's New

Salesforce Connector

Salesforce Connector MP supports relation synchronization from Salesforce to Reltio when a child object in Salesforce contains a reference to a parent object. For example, Contact is a child of Account.

To enable this relation functionality, the following steps must be performed:

  • Child object mapping must be configured to use relations.
  • Relations array must be added to a relation object.

One Salesforce object can have several relations to other Salesforce objects (Child Sobject has separate fields with reference to parent for each relationship). For more information, see Relation Synchronization from Salesforce to Reltio Platform .

Console

  • You can now identify a column from the input file as a Record Identifier. When you download match results, the record identifier allows you to identify the input records that matched with the records in the tenant. The column that you identify as the record identifier must have unique values. By default, the first column is selected as the Record Identifier. For more information, see Mapping File Columns to Attributes and Exporting Match Results.
  • When downloading files from Export in Reltio Console, we have reduced the multiple-file download size from 2 GB to 1 GB. This is to avoid timeout restrictions that some users were experiencing with the larger download file sizes. For more details, see Downloading Exported Output Files.

Reltio Platform

  • Events API is used to get event statistics and to retrieve event objects by filters. This API allows you to get statistics of detailed synchronization process such as a number of creates, updates, deletes, and merges events per given period.
  • Added support for auto-generation of IDs for nested attributes. For more information, see Auto-generated IDs for Simple and Nested Attributes.

Fixes

Console

While using Reltio Console, users were unable to select or deselect (using the Select All/Unselect All option) all the selected files at a time. In addition, they were unable to download all the exported files even though the file size did not reach the maximum limit. These issues are now fixed. The issue related to Select All/Unselect All files is solved by correcting the incorrect referencing of a variable to the data table. To resolve the download issue, the payload limit of 100 KB is now configured to 1 MB and the download limit is restricted to 1 GB.

Hub

PostalCode and Zip5 were not getting exported from Hub (with or without the Operational Value (OV) value option) either in CSV or JSON format. This issue is now fixed by correcting the logic associated while parsing nested attributes.

Search

The Search filter was built incorrectly with blank space in the URL. This issue is now resolved by fixing the Search filter.

Data Cleansing

Phone Cleanser was cleansing the Country correctly but the CountryCode were not cleansed as expected. This issue occurred because the area code was not considered if an international code had more than one region. This issue is now fixed by including a validation to check region codes when more than one countries are present in the same international code.

Reltio Platform

  • When a relation was created as a refRelation of an entity, either startObjectCrosswalks or endObjectCrosswalks (depending on the direction of relationship) were found to be empty. This issue is now fixed.
  • While unmerging entities created in Reltio, nested attributes containing multiple crosswalks were not re-assembled properly. As a result, after unmerging the entities, nested attributes appeared as split. This issue is now fixed by including an analysis to unmerge the single nested attributes irrespective of the number of crosswalks that the entity contained.
  • Reltio jobs were failing and the search was not working as expected because multiple events were generated for large objects by the background processes. This issue is now fixed by including additional limits for background processes to avoid such overload.
  • Inconsistency was caused as the tenant storage was configured with an incorrect cluster name and host name. This issue is fixed by enabling the hosts list validation that ensures all the specified hosts must belong to the same cluster.

Match and Merge

  • Cleansers from different match rules affected each other and led to the unexpected merging of entities. This issue is now fixed.
  • The CustomComparator did not support the fuzzy comparators accurately. This issue is now fixed.

Release Notes for 2019.3.0.15 Emergency Bug Fix Release - Jan 24, 2020

Fixes

Workflow

While approving Data Change Requests (DCRs), the inserted reference attribute changes were applied to the referenced entity defined by objectUri. This resulted in the complete overriding of the attributes for the default Reltio crosswalk. This issue is now resolved and reference entity attributes remain unchanged.

Release Notes for 2019.3.0.14 Weekly Release - Jan 22, 2020

What's New

Salesforce Connector

The useBulkFrom parameter is added to the Connector profile to allow you to post bulk objects to Salesforce in a single request by using the Bulk API. For more information, see Tenant Profile Configuration - Get all accessible profiles for a tenant, Get Salesforce Connector Profile Configurations, and Set Salesforce Connector Profile Configurations.

Hub

This enhancement introduces a change while displaying Potential Matches. With this change, the Potential Matches perspective and Potential Matches facets display only the active entities in the Potential Matches list. For more information, see Potential Matches Perspective. To view Potential Matches with different activeness status of the matched entity, you can use the Potential Matches APIs. Added a new parameter activeness for the Potential Matches APIs with the _matches, _notMatch, and _transitiveMatches endpoints. By default, the Potential Matches APIs list all the entities. You can set the activeness parameter to active to view only active entities in the response for the potential matches API. For more information, see Potential Matches API.

Console

  • Data Loader is enhanced to include the Enable Data Change Request parameter, which when enabled creates Data Change Requests (DCRs) for entities in the input file. The entities are loaded into the tenant only when the corresponding DCR is approved. For more information, see Specifying Record Update Preference.
  • If some values are missing in one of the mapped columns from the input file, new options are introduced to set default values or remove values from the tenant during partial override. For more information, see Mapping File Columns to Attributes.

Fixes

Match and Merge

The reindexPotentialMatches and mergeAutoMatches values were missing in the RebuildMatchTableV2Task JSON. This issue occurred as reindexPotentialMatches and mergeAutoMatches were not saved to a RebuildMatchTableV2Task state. This issue is now resolved.

Export

While exporting relations using the Export API, the Select filter was not working as this filter was disabled for export. This issue is now resolved.

Console

  • Data Loader jobs were stuck in a queue and could not pass the MAPPING_PROCESSING. As a result, longer time was expected in transforming the data from a CSV to JSON. This issue is fixed by enabling parallel processing.
  • While uploading a file from the local file system, the file size must not exceed 50MB. For more information, see Loading Entities into a Tenant.
  • Reltio Console could not find some RDM tenants as the number of returned tenant identifiers by the RDM API was limited to 1000. This issue is fixed by increasing the limit to 10000.

Reltio Platform

  • The POST: https://361.reltio.com/reltio/api/Eab8HWG8cdQfx4z/entities/_scan?filter=equals(id,'7H888is')&select=type,uri,updatedTime&max=1 call and the "GET: https://361.reltio.com/reltio/api/Eab8HWG8cdQfx4z/entities/7H888is?select=uri,type,updatedTime call returned mismatched values for UpdatedTime. This issue is now fixed.
  • After completion of data load, entity count dropped continuously and the entities were eventually deleted. This issue occurred due to incomplete clean-up in the tenant index. In addition, the consistency check created incorrect records in the ENTITIES table while adjusting the object version between the primary and secondary storage. This issue is now resolved.
  • The logic to retry match events did not consider a persistent merge problem. As per the issue, the entities could not be merged as the maxCrosswalks limitation was reached. To reduce the number of hanging events, the retry logic had to skip similar issues. A large number of retries led to the tracing of multiple match events and eventually to Out of Memory (OOM) errors. This issue is now fixed.

Release Notes for 2019.3.0.13 Weekly Release - Jan 15, 2020

What's New

Salesforce Connector

Fixes

Console

Dataloader failed to load data and displayed the reason, Failed to parse the JSON. This issue is now resolved.

Workflow

While approving Data Change Requests (DCRs), the inserted reference attribute changes were not applied to the referenced entity when crosswalk information was not defined. This happened when the Reltio crosswalk was created by default. This issue is now resolved.

Data Cleansing

By default, US appeared as the Country Code irrespective of the actual Country Code of a profile. This issue occurred because after cleansing the lookups were not resolved and the attributes of RDM values were Null. Now, the lookups are resolved before cleansing and the RDM values are considered for attributes.

Match and Merge

Multiple entities were not merged because the match tokens were not in sync with the version of the match document. This issue is now resolved by not allowing the match tokens to be written for an outdated match document.

Export

Spark export failed for AWS because Spark checked the existing path of all sub-directories for a custom S3. This issue is now fixed by modifying the S3 client to check only the directory where the index data is stored.

Reltio Platform

  • When performing bulk update, if the IsClosed attribute was updated with Boolean value (true or false) for one crosswalk, the other values from the associated crosswalks were also getting updated and ignored automatically. This issue is now resolved.
  • Collaboration Service works as expected and loads in a reasonable amount of time.
  • Dataload JAR failed to load data during the first run but loaded the data subsequently. This issue is now resolved.
  • CopyAutoMatchesTask requires fully up-to-date tenant configuration including the information about the rebuildInBackground mode set in the rebuildmatchtable request. The up-to-date tenant configuration was not gathered by some tasks and hence the tasks failed. This issue is now fixed by requesting for updates before the task is executed.
  • One of the production tenants was experiencing timeouts and high response time when reading entities. The database that hosts the System Keyspace namespace was overloaded and that caused 100% CPU consumption on multiple database nodes. As a result, the GET /entities/<id> requests were not processed when read from the PERMISSIONS table (from the System Keyspace). This issue is now resolved by reducing the load on the PERMISSIONS table.

Release Notes for 2019.3.0.12 Weekly Release - Dec 30, 2019

What's New

Workflow

Added the ability to partially reject the attributes in a Data Change Request for new entities and relationships from the Inbox UI. For more information, see About Data Change Request.

Release Notes for 2019.3.0.11 Weekly Release - Dec 26, 2019

What's New

Workflow

A new Workflow API is introduced that allows you to modify existing changes or add additional changes to a Data Change Request (DCR).

Reltio Platform

The behavior of partial override functionality is enhanced to ensure that you can update an existing nested attribute value and also add new values to a nested attribute. This ensures that you are not required to load the entire profile into the tenant. The enhanced behavior is also supported to update or append values at the sub-attribute level. For more information, see enableNestedPartialOverride Option and keyAttributesURI Option.

Console

  • You can now assign a default value to any attribute (simple or nested) without mapping the attribute to any column of the input file. For more information, see Mapping File Columns to Attributes.
  • In Data Loader, you can now delete an existing mapping if it is not used by any project. For more information, see Deleting an Existing Mapping.

Fixes

Export

The RIQ Data Export for HCP through Export V2 was affected with duplicate key issue and encountered the EXPORT_ANALYTICS_REMOTE_ERROR error. This issue is now resolved.

Salesforce Connector

Empty nested attributes were created if SF value was equal to null and was mapped to nested attributes with some pre-existing sub-attributes. To resolve this issue, an expression is added to compare the SF value with null (for example, compare(null,null,someValue)). If the value is not equal to null, then the SF value is considered as constant. You can also create additional comparison for constant value and set them by using conditions.

Reltio Platform

  • During serialization (conversion from Java Object to JSON) of entities that contained self-reference attributes, some connections between crosswalks and attributes were lost. This caused incorrect association of attributes and crosswalks when an entity was returned from LCA. This issue is now resolved. However, the fix does not affect the already corrupted entities and those corrupted entities must be reloaded.
  • Multiple entities could not be edited if the entities contained ov:false value at the parent level and the sub-attributes had OV value marked as required:true. As a result, validation errors were encountered. To fix this issue, an additional logic is included to check if an entity at the parent level is ignored or contains ov:false so that the OV value is ignored during validation.
  • A user with the Local Data Steward's role could create the Local Medicinal Product (LMP) entity without a Global Medicinal Product (GMP) value inside it. However, the user could not create a LMP entity if the GMP value was inside it. This issue occurred as the logic for validation permissions for creating a new entity with a reference entity was different for the Validation Service and the Reltio Platform. With this release, this issue is now fixed by correcting the logic.
  • When trying to update a sub-attribute of the Address reference attribute, the entities/.../_update endpoint request failed to run as the relationship had a non-existent entity URI in the endObject. When removing the loser entity, it was observed that the CollapseLosers task failed to update the start or end object field inside a relationship as the winner entity. This issue is now resolved.
  • In multiple scenarios, the Data Import Job failed without any error description. This issue was an intermittent exception that occurred due to Eventual Consistency in S3. This issue is now resolved.
  • JAR deployment per tenant failed due to duplicate keys. This issue is now resolved.

Release Notes for 2019.3.0.10 Weekly Release - Dec 23, 2019

Fixes

Reltio Platform

The GET /entities call when specified in the permission configuration as /permissions/{tenant}, could not find entities that had multiple missing attribute values (for example, attributes.Source) because the data security filter in the configuration did not match the attribute condition (for example, not equals(attributes.Source,'EVEREST')). However, those entities were found using the GET id /entities/{entityId} call.With this release the issue is fixed for the GET /entities call. Now, the data security filter is applied for both, searching and retrieving entities by id. As a result, entities with missing attribute values now satisfy the not equals condition.

Release Notes for 2019.3.0.9 Weekly Release - Dec 19, 2019

What's New

Salesforce Connector

  • A new parameter called OwnerId is introduced in SBC import request that allows you to set the owner for the Salesforce Object. For more information, see Import Entity.
  • The useFuzzy option is added to the SBC API search endpoint to perform fuzzy search for String attributes only in Reltio. With the fuzzy search option, you can even search for values with typing errors and SBC returns suggestions of existing records based on the fuzzy search logic. For more information, see Fuzzy Filtering .

Fixes

Search

Entities that were not potential matches were also appearing in the search result. This issue occurred when simultaneous changes were made in potential matches for an object. This issue is now resolved. In addition to this, the PotentialMatchesCassandraESConsistencyTask is improved to fix the existing inconsistencies.

Hub

The Graphs and Hierarchy facets did not display any information even if a user role had read access for all the entities/relations/graph/hierarchies. This issue occurred as the request permissions were calculated for objects without considering their reference attributes. As a result a user with permissions defined by a filter applied to reference attributes, encountered a security exception error. This issue is now resolved by changing the object type for calculating permissions to include reference attributes.

Reltio Platform

  • After a Cumulative Update for changes is made in the Hub, the SingleAttributeUpdateDates section got updated for attributes that were not listed or related to the Cumulative Update. This issue is now resolved as the SingleAttributeUpdateDates field for attributes that were not updated is now restricted from Cumulative Update.
  • The Ignored flag for an attribute is set by the value itself and not by crosswalk or URI value. For example, if ignore value is 123 then the values from all crosswalks that are 123, are ignored. While performing bulk update with an already ignored value, it was observed that for entities (with a new value) the new value was ignored automatically. This issue is resolved by removing the Ignored flag from the old value.
  • Tracing of failed match events consumed too much memory in case of multiple matches. This issue is resolved by adding limitations for the amount of information collected for the match events.

Release Notes for 2019.3.0.8 Weekly Release - Dec 11, 2019

What's New

Reltio Platform

After you remove a relation, the updatedTime and updatedBy properties of the main entity are updated accordingly. The main entity is the startObject of the relation. For more information, see Entities API.

Fixes

Reltio Platform

In some scenarios, multi-nested address values were not getting merged. This issue occurred as the Reltio platform failed to collapse the multi-nested attributes and then check the matchFieldUri field for the nested attributes. This issue is now resolved.

DTSS

If a crosswalk was end-dated to a future date and the Golden Record Crosswalk (GCR) was active in the DTSS subscription, then the crosswalk could not be imported. With this release, this issue is now resolved.

Search

When trying to group survivorship rules as per Country and Source, appropriate results were not displayed. However, on removing the Country filter, results were displayed. The issue occurred as the search engine performed the search based on an invalid survivorship group that led to no results.

With this release, this issue is resolved by fixing the search engine to use the survivorship group in the correct manner.

Export

A temporary folder is used by the export job to place files before copying to the final destination folder. We made a change to ensure the working directory name is unique for each export task.

Console

Tenant Management used hazelcast.environment parameter from system configuration to filter metrics and show the information for a selected tenant and environment. However, in one of the previous releases, the metrics were moved from the hazelcast.environment parameter to another configuration parameter, reltio.environment.name. As a result, the Tenant Management UI did not display the queued information since the date of that change. With this release, this issue is now resolved by removing the option to filter metrics by environment. Metrics are now filtered using tenant id only as the ID is a unique value. Tenant Management now displays the queued information even for the date when the issue was reported.

Hub

A Saved Search is created on the Hub with the current date or time as the Last Modified Date. However, on updating the Saved Search, the Last Modified Date field was not updated. This issue is now resolved.

Salesforce Connector

When the lookupCode function was enabled, the Search Before Create (SBC) by attribute values was not working as expected. This issue occurred as SBC did not process the lookupCode() function in the correct manner. This issue is now resolved.

Release Notes for 2019.3.0.7 Weekly Release - Dec 04, 2019

What's New

Match and Merge

The behavior of the matched field in the response for _verifyMatches is changed. Now, the intersection of tokens between two entities is considered to evaluate if the two entities are matched or not. The matched field indicates that tokens have intersections and the legacy behavior is used for matching. This ensures that there is a better understanding of why entities matched or not. To preserve the legacy behavior, a new field called matchBydocuments is introduced that identifies matches by using the legacy matching behavior only.

Fixes

DTSS

When manually syncing data between Data Tenant (DT) and Customer Tenant (CT), if an error occurred during entity processing, an infinite loop was created by the three threads, scanCtEntities, processEntities, and uploadEntities. Due to the infinite loop, the entity read cursor did not proceed to sync the rest of the batches. With this release, this issue is fixed. Now, if an error occurs while syncing data between DT and CT, the entity read cursor skips the error and proceeds to the next batch of entities.

Hub

Potential matches, when searched by Source, did not return the expected result. This issue occurred because Hub used the source system label as the search criteria on the Potential Matches screen. If a source system label was not the same as the source system abbreviation, the search did not return the expected results. As a result, entities could not be filtered as per the source system. With this release, this issue is resolved. Hub now uses the source system abbreviation as part of the search criteria, similar to how it behaves on the Search screen.

Match and Merge

The ExactorAllNull operator for match rules did not work as expected. This issue occurred due to the incorrect implementation of cross-nested cleansing for nested attributes, which affected the regular cleansing for nested attributes. With this release, we have disabled cross-nested cleansing for nested attributes because even though this feature is not used, it brings in a redundancy.

Reltio Data Science

URLs for new environments were not added to UI RIQ configuration. With this release, the URLs are now added.

Search

Search results were appearing with a greater delay than usual. The longer processing was caused by an L3 configuration update (done by customer) that affected many matches. It is recommended to wait until the processing finishes (internal queues should be empty).

Reltio Platform

  • A validation error appeared notifying that RDM canonical value mapping was not found for reference attributes. This issue occurred as the RDM transcoding request was incorrectly built for reference attribute, wherein it had reference to the same type as the main entity. This issue is now resolved.
  • Entities appearing as suspected matches did not show any potential matches. This issue occurred because even though an entity was deleted, its potential match record was not deleted.

    With this release, the Consistency Check task has been improved to fix such issues.

  • Two hidden attributes namely, SalesforceOpportunityNumber and SalesforceOpportunityKey were added to the nested attribute ProducerCode, of the Contact entity. When trying to retrieve these attributes by using an API call, the attributes were hidden on the Hub as expected. However, the attributes and their values appeared on the Potential Matches screen.

    With this release this issue is resolved by filtering out the hidden attributes from the Potential Matches screen.

Salesforce Connector

  • When working with Salesforce (SF) org, you can create new record types for objects in SF. The SF objects are eventually mapped to the Reltio entities. It is also possible to map one Reltio entityType to one sObject but with different recordTypes. In order to resolve these kind of mappings, there was a need to compare with multiple values.

    With this release, we have introduced the functionality to compare with multiple values in a mapping statement. Now, you can compare for n number of recordTypes to ensure proper mapping is done between SF objects (with recordTypes) and Reltio entities.

  • In the Salesforce Connector V2, constant values were not validated by the mapping validator. Also, the expressions in the connector were not processed from Reltio to Salesforce. As a result, the recordTypeId field could not be set. With this release, this issue is resolved as the validator now validates constants and the connector processes the expressions from Reltio to Salesforce.
  • SQS queues were processed slowly when entities did not exist. With this release, the Salesforce Connector now loads 10 entities at a time when reading the SQS queue. Also, the connector does not retry event processing if an entity is missing in a tenant.

Release Notes for 2019.3.0.6 Weekly Release - Nov 20, 2019

What's New

Reltio Platform

We fixed a bug that was preventing the extraction of reference attributes in relationships, if they were present at the end object of the relationship. Now, users have the ability to extract reference attributes in a relationship (regardless of direction).

Fixes

Match and Merge

The entities identified as potential matches by using the _verifyMatches API were not merging due to inconsistency in the match storage. However, running the reindexing task resolved the issue sometimes. With this release, the _verifyMatches API is enhanced by setting the foundInMatchTables field to true to identify the inconsistencies.

Life Cycle Actions

Entity and Relation data are serialized (converted from Java Object to JSON) before sending the data to the LCA services. However, deserialization (conversion from JSON to Java Object) is done only after the LCA service returns a response. The LCA deserializer only considers the start and end object objectURIs and not the crosswalks. Therefore, after the LCA call for the beforeSave hook, the start and end crosswalk values were not available. The issue is resolved as now the crosswalk is also considered during the LCA call.

Salesforce Connector

When you deleted a nested attribute in Hub, it was not getting deleted in Salesforce. Now, to ensure that a nested attribute deleted in Reltio is also deleted in Salesforce, mark the nested attribute as InActive. The connector uses this InActive flag and deletes the nested attribute in Salesforce, as per the expression mentioned in the deleteExpression parameter.

Reltio Platform

  • Multiple customers with SSO authentication could not use the Export UI and received the 400 error. These errors were caused by changes made in the NodeJS in the form of heavy requests (requests with long headers). The requests were incorrectly parsed and eventually were rejected with the 400 error because there was a hardcoded parameter in the http_parser library (80 KB), which was changed to 8 KB.

    When SSO authenticated users with long access and refresh tokens exceeded the 8 KB limit, the requests failed.

    However, as the http_parser library parameter is configurable, the value of the hardcoded parameter is changed to 80 KB.

  • When loading data to a tenant, the matching queue went to the Yellow state but the queue size did not reduce. This issue occurred due to the incorrect version implementation for the match storage, which started to have impact only after rolling out the Version Check feature in the 2019.3 Release. With this release, this issue is now resolved.

Export

  • Reltio operates a multi-tenant managed service. This means our infrastructure resources are shared between multiple tenants, which helps us operate the service in an efficient manner. We use an internal queuing system to schedule export requests when resources are busy serving other requests. This helps us provide a more resilient service, with fairer resource allocation across tenants.

    We use auto-scaling to increase capacity when workloads reach pre-defined threshold. Autoscaling can take several minutes per unit of additional capacity. In export version 1, the parallelization level (number of distributed task parts) is normally specified in the request for large jobs. With export version 2, the parallelization level is computed automatically based on the job size.

  • Export jobs with the Date filter had different results for the number of attributes or entities count. On selecting the start(Since) and end(To) date, the time must start with 00:00:00 and end with 23:59:59. However, the same was not happening as the request body was constructed with the incorrect start and end time (hh:mm:ss). With this release, this issue is now resolved.

Release Notes for 2019.3.0.5 Weekly Release - Nov 13, 2019

What's New

Console

With this enhancement in the Export UI, now there is no restriction on the file size when you download a single export output file. However, when you download multiple files, the total file size is limited to 2 GB.

Salesforce Connector

  • Added support for mapping one entity type to two or more Salesforce orgs. With this, when value in entity is changed, the same will be synchronized to all the mapped Salesforce orgs.
  • Added support for default Salesforce object merge by calling the Salesforce merge service when entities are merged in Reltio.
  • Added support for mapping Reltio reference attributes to the built-in address of accounts or separate custom objects. To enable the mapping of reference attributes, specify the reference entity type and relationship type in the to_salesforce and the to_reltio mappings.

Fixes

Match and Merge

The relevance based match rule returned 1 instead of 0 even though the city and state values between profiles did not match. This occurred as non-OV values were considered null when used in the exactOrNull operand. Now, if the useOvOnly parameter is set to true, and the attribute has OV and non-OV values, then the nested non-OV values are skipped.

Workflow

When opening the Inbox, Team, or the All Folder menus in Inbox, the list of Data Change Requests (DCRs) were not displayed. This issue occurred as the default timeout from the Inbox side was 120 seconds. It was observed that the heavy requests were failing by 502 seconds. To resolve this issue, now the timeout from the UI side is removed. However, the timeout for heavy requests on the API side continues to exist and requests can be canceled by the API only.

Reltio Platform

  • Bundle entities of export jobs could not be downloaded from Reltio Console. This issue occurred as the UI request with long HTTP was rejected by the Tomcat server with the 400 error. The headers are long as they contain cookies with very long access and refresh tokens generated by the SSO auth service. As these cookies are not required by the Core API hosted on Tomcat, the cookies are removed before sending the request.
  • Two relations under the same crosswalk had different end objects as the API simultaneously recreated the removed crosswalk. This issue is now resolved.
  • Existing relations were not displayed on the Hierarchy facet. This issue occurred as the API skipped some relations while building graphs by filtering them using non-resolved end objects. The API is fixed to ensure it does not skip the relations.

Release Notes for 2019.3.0.4 HotFix - Nov 07, 2019

What's New

Console

Now, in Reltio Console, an appropriate error message is displayed when a user tries to upload a file, which is of a version or format that is not supported.

Fixes

Match and Merge

New parameter resolveLookups is added to the External Match API. When the resolveLookups parameter is set to True, the lookups are resolved for entities in the input file. The default value is False. For more information, see External Match API.

Export

We fixed a bug in the calculation of the number of resources required to execute all the requested export tasks in parallel.

Reference Data Management

When using the REST API request GET /lookups/{code}/{codeValue}, the URLs were not validated. This impacted the user-experience because no informative message was displayed for invalid requests. Now, an additional check is included for validation to ensure that an error message is displayed if the {code} or {codeValue} is not available.

Reltio Platform

  • The country description for Country Codes did not display all the values from RDM due to the caching mechanism. As a result, when performing an advanced search for dependent lookups, only the first 50 values were added to the cache. For subsequent searches, the same cache was used for the searched facets and no additional requests were made to resolve the new lookup codes. As a result, the Country Code facet did not display all the values.

    This issue is fixed by including additional requests for the non-resolved lookups that were added for the Attribute facet.

  • When creating a new entity, the default values for the attributes are populated. An issue was observed as the Validation Service was timed out and the default values for the attributes could not be saved or recovered at all. This issue occurred due to the simultaneous updates that were created for a temporary entity structure (attribute view). The temporary entity structure is prepared while creating a new entity during the entity creation process, after which the default values are populated.

    This issue is now resolved by including an additional logic to prevent simultaneous updates.

  • Data Loader APIs were creating duplicate periodic tasks because the acknowledgment was not received within the stipulated time of one minute. Now, the acknowledgment time limit is increased to 10 minutes that ensures no duplication of the periodic tasks.
  • When an attribute was pinned, the OV flag was set to true without checking the crosswalk end date. Now, if an attribute is pinned and the crosswalk is end dated, then the OV flag is set to false.

Console

  • When loading data through the Data Loader application, the Dataload task was taking considerably longer time to complete. This occurred because several HTTP connections were opened to read error files that resulted in a timeout.

    Now, HTTP connections are opened only when the error files can be read and closed immediately after the read operation.

  • When loading data through the Data Loader application, Phone Numbers with attributes Type and Value were not displayed properly on the Reltio UI. This occurred due to an incorrect processing of mapping. This issue is now fixed.

Release Notes for 2019.3.0.3 HotFix - Oct 30, 2019

What's New

Export

We added support for the ZIP archive format with compression for the Spark based export strategy in export version 2. With this change, both the ZIP and GZIP file compression formats are supported for Spark based export. A tenant configuration property is used to define which compression format to use. For more information see, Export Version 2.

DTSS

The updated crosswalk for a relationship in Data Tenant (DT) was not getting synced with the Customer Tenant (CT). For example, if you set a future date as a delete date for a crosswalk in DT, the crosswalk was updated in the customer tenant with the delete date as the system date. This crosswalk update synchroniation from DT ro CT is enabled now. For more information, see DTSS Configuration Procedures.

Fixes

Hub

The Specialty drop-down list must be populated based on the selected Specialty Type and Country. However, the Specialty drop-down list was displaying all values instead of values specific to the selected Specialty Type and Country. This is fixed to ensure that the Specialty drop-down is populated with values based on the selected Specialty Type and Country.

Match and Merge

An issue was found in the incremental match thread that resulted in the double-processing for the same matches. As a result, the match profile count on the dashboard increased significantly. This is fixed now.

Export

Spark export was taking considerably longer to complete the export. This was because the winner entity link resolve was causing performance degradation on tenants with a large number of relation types. To ensure that there is no performance degradation, the resolveMergedEntities flag, which is now enabled to support Spark export, is disabled by default. As a result, when relations are exported as part of the export, the export results will contain links to only loser entities, but there will be no performance degradation.

Console

From the Export UI, if you triggered the export job for entities with the operational value (OV) set to false, the job was getting results for OV set to true. This occurred because for export version 2 (v2), the OV value was passed through the querystring (URL) instead of the request body. This is fixed now to ensure that the export job run from the Export UI with OV set to false or true works correctly.

Reltio Platform

  • If a merge occurred, it was present in the activity log, however, the merge was missing from the history tree in the profile in Hub. The history of merge events was written synchronously without a queue but there was a possibility of events being missed. As a result, the merge tree was not getting updated. Now, a new processor for the match queue is added to write the history of merge events and update the tree.
  • When searching for Person ID by uploading a file in the Advanced Search, no search results were displayed. This was because the generated S3 link had an invalid SSL certificate with periods (.) in the bucket name and the platform rejected this link. This is fixed now.
  • DCRs were not appearing in the Reltio Inbox. This occurred because the UI request with long HTTP headers was rejected by the Tomcat server with the 400 error. The headers are long because they contain cookies with very long access and refresh tokens generated by the SSO auth service. As these cookies are not required by the Core API hosted on Tomcat, the cookies are removed before sending the request.
  • The Sequential Generator was assigning the same cmdid to multiple crosswalks linked to different golden records and contributed by different sources. When Hazelcast was unstable, the Sequential Generator had some issues in the algorithm that resulted in duplicate values being generated. This is fixed now.
  • The activity log was getting updated repeatedly for potential matches even though there were no new potential matches. The ENTITY_UPDATED match service event was unnecessarily getting repeated in case of match document and entity version mismatch. As a result, the ENTITIES_MATCHES_CHANGED CRUD event was triggered repeatedly thereby updating the activity log. Now, the ENTITIES_MATCHES_CHANGED is triggered only if there are merges or potential matches identified

Release Notes for 2019.3.0.2 HotFix - Oct 23, 2019

What's New

Reltio Data Science

RIQ Export saves Spark dataframe into a parquet file. This procedure requires the column names to be in the proper format to avoid errors. Therefore, if the analytics attribute contained blank spaces in the name field value, an error was reported. With this release, the naming of the Spark dataframe column is enhanced by allowing the use of attribute name from the uri field.

Console

While loading data, in the Select Mapping window, you can now sort existing mappings by using the Created by or Created date options. For more information, see Mapping File Columns to Attributes.

Reltio Platform

  • Added the ability to limit the total number of export tasks running in parallel at any given point in time. The limit is 20 by default, but you can override the limit in the tenant level configuration.
  • Added a storageType property with a value of S3 (for AWS S3 Bucket), GCS (for Google Cloud Storage), or AZURE (for Azure Blob Storage) in the response to Export GET Task status API calls. The storageType property is listed under the currentState JSON object. For more information, see Export Tasks Management API.
  • Prevented publishing of events for profiles where the profile definition is not found in the tenant L3 configuration. Such events should not be processed by the consumer until configuration is corrected. This also avoids expending unnecessarily infrastructure resources on such events.

Fixes

Workflow

The POST /assignee call did not assign the task to users defined with necessary permissions for group level and restricted permissions for roles at the same time. This issue occurred as the default value for resolveGroups was false. With this release, the workflow is fixed to use resolveGroups=true.

Match and Merge

When using multi in the match rule, matching was not taking place as expected. This issue occurred as the OV values were not handled properly. With this release, this issue is resolved by considering OV values only if the match rule is marked as useOvOnly=true.

Export

In Export Version 2, the file extension of multiple export files was missing. With this release, this issue is now fixed. For more information, see Export Version 2.

Reltio Platform

  • Multiple entities with the same URI appeared as part of search results. This issue occurred due to the absence of retries in cases where the processors did not process the events successfully. With this release, this issue is resolved by properly handling and logging the temporary failures for events.
  • An issue was reported wherein the value of the Reltio crosswalk was not equal to the contributor (usually Id of the objects) it belongs to. With this release, this issue is resolved by adjusting the value of the Reltio crosswalk to make it equal to the contributor Id.
  • The logic to identify whether an event must be repeated or not was not working correctly. As a result, it repeated events even if all the matches were already merged to another entity. With this release, the logic is modified to properly handle events for which the matches are already merged or changed. In addition, additional logs are included to simplify the investigation when an event is repeated multiple times.
  • In the 2019.3 release we made the user email address field mandatory for Single Sign-On (SSO) authentication. However, there are multiple customers who do not use this field for SSO. As a result, login to Reltio failed for customers where the field was not set. We have now reverted the field from mandatory back to optional which has resolved the issue.
  • Crosswalks of some entities displayed unexpected data related to the relation URI of the reference attributes. With this release, this issue is resolved by displaying such reference attributes in the list of attributes for the entity

Release Notes for 2019.3.0.1 HotFix - Oct 18, 2019

What's New

Reltio Platform

In 2019.3, Reference Data Management (RDM) was enhanced to provide the ability to search for RDM lookup errors in Hub without making configuration changes. As a part of this change, RDM transcoding failures were displayed as Warnings on the Profile screens in the Read and Edit modes. The end-users and our customers did not expect to see these RDM validation warnings.This enhancement ensures that when you open entities with RDM errors, the validation warnings are not displayed. To see these warning messages, you must submit a Support Ticket. For more details, see Editing Profiles.

Fixes

Export

Reltio Console switched internally to use a standard export API. However, for tenants that were using export version 1 (v1), the logic to select the optimized export strategy was missing in the export service. In this fix, the missing logic is added to the export service. For more information, see Export Version 2.

Reltio Platform

  • When fetching entities by crosswalk using the GET API, incorrect entities were returned. After this was fixed, getting entities by crosswalk returned only entities that had a Reltio crosswalk. Now this behavior is changed. If an entity does not have any default crosswalk (Reltio) and the searchable crosswalk does not belong to any other entity, then the entity is returned by ID. For more information, see Get Entity by Crosswalk.
  • The data exported from Hub was different from the data exported from Reltio Console. The difference appeared for lookup values inside reference attributes. Hub exported value, lookupCode, and lookupRawValue for reference attributes. However, Reltio Console exported only the lookup value inside reference attributes. The difference occurred because Reltio Console did not resolve lookups for reference attributes when useOnlyOvValuesInReferencedEntities was set to True. Now, the Resolve Lookups service is changed to ensure that Reltio Console exports value, lookupCode, and lookupRawValue for reference attributes.