TOLERANT Match Release 11.0

2022/05/02

Notices

  • CLI-Parameters for config files and project ids have been harmonised:
    • Project ID: projectid
    • Configuration file: configfile

    This affects

    • MatchAnonymizer
    • MatchClusterTool
    • MatchDataProfiler
    • MatchSynonymAndNormalizerTestCLI
    • MatchAdminCLI
    • PostDataProfilerCLI
    • PostSynonymAndNormalizerTestCLI
    • PostAdminCLI
    • NameAdminCLI
    • BankAdminCLI
    • CarAdminCLI
    • MoveAdminCLI
  • Logging rollover strategy was corrected for rolloverMode TIME. The rollover strategy will use the file name for deleting old files. The file names must therfore be chosen in a way that no file name is a prefix of the file name of another log element.
  • The software now requires a minumum version of GLIBC 2.17 for Linux operating systems.
  • Completely reworked the GUIs of all products with the Angular framework. There is now a new entry point, the “Portal”, accesible via ://:, which replaces the former static html entry page. From there, multiple installations of all TOLERANT products can be managed. For additional information please refer to the manual.
  • A new version of the service has been added (V3). This REST service interface defines endpoints for all clients (product GUIs and third party clients) and is used for administrative purposes as well as functional, productspecific calls.
    The shipped Java API will use the V2 version for compatibility purposes until further notice.
    All relevant endpoints can be looked up via the GUI Information menu in the portal, or directly by calling ://:/Service/swagger.
    The V1 Version is now considered deprecated and will be removed with version 12.0 of the product.
  • Default behaviour of generated H2 databases changed (since 10.0). The created databases now consist of a different filename extension (mv.db). To be compatible with older versione, one must set the parameter „MVCC=false“ in the JDBC URL.
  • The uniqueness check for id attributes in configuration elements is now case insensitive. After an upgrade this may lead to initial configuration check errors on startup which can be resolved by making the ids unique. This may primarily occur in configurations that were derived from our shipped examples in the element inputFieldmap .
  • Because of substantial internal changes, an upgrade from earlier versions to version 11 is no longer supported and a new installation is needed. From version 11 upwards to higher versions, the upgrade functionality will be available again.
  • Since the ojdbc driver is deprecated, TOLERANT products will now use the UCP datasource per default. You can still opt in to use the old driver by setting the environment variable OracleCache_UseUCP to false.

New Product Features

Configuration and Administration GUI

  • Added new GUIs for all products.
  • The GUI help section now allows to see the xml element- and attributs names (in the configfile) of the current field or view one is working on.

Batch Wizard GUI

  • Batch Wizard has been updated to most recent version of Angular.

Batch

  • The partitions are now processed in the reverse order of their size, in order to improve the performance of batch processing.

Service

  • Added new REST service version 3.
  • Improved bulk service handling: Frequency of the bulk cleaner job can now be influenced by the configuration parameters responseStorageTime and responseDoneStorageTime.

General

  • The bundled JRE was upgraded to the latest Java security patch (Java 8 Update 322).
  • A connection to an oracle database can now be established using UCP as datasource.
  • A new attribute tokenIgnoreMultiple was added in matchProfileField. This new attribute can be used to control whether duplicated tokens are to be considered in the field scoring.
  • Added additional features to support.bat/sh tool:
    • Information about filestructure and permissions of the installation folder were added. By using the parameter -depth [number] the output can be limited.
    • The parameter -logsize [MB] was added to limit the size of the logfiles that would be included in the support.zip,
    • The parameter -logrange [hours] was added to limit the timespan of the logfiles that would be included in the support.zip.

    For additional information regarding the usage please refer to the manual.

  • Temporarily removed upgrade functionality in installer.
  • Extended error messages for file checks (normalizer or synonym files).

Fixed Bugs

Batch

  • TPR-9033: Fixed an issue that caused fields with the type date to not function properly in case of batch runs.

Service

  • TPR-6765: Prevent data loss in mode MASTER_FORCE_OVERWIRTE under specific conditions in cluster mode.
  • TPR-8881: Possible race condition problem in the check for duplicated keys during the initial load was fixed.
  • TPR-9319: It is now possible to add more than 50 serviceInputFields in the configuration.
  • TPR-9362: Match service now uses the HTTP Method ‚POST‘ for insert requests and the HTTP Method ‚PUT‘ for update requests.
  • TPR-9756: The handling of aggregation partitioning was corrected to prevent the service from freezing in some boundary cases.
  • TPR-9987: Reconfiguring the service now handles deleted projects correctly.

API

  • TPR-8882: The C-API method matchGetResultRuleName now returns the configured rule name as string.
  • TPR-9757: Errors when using JAPI with injected loggers are now logged properly.

General

  • TPR-6211: Fixed inconsistencies in CLI tool usages.
  • TPR-6473: Removed unnecessary files from installer bundle.
  • TPR-6599: Rollover strategy was corrected for rolloverMode TIME.
  • TPR-6874: The software now requires a minumum version of GLIBC 2.17 for Linux operating systems.
  • TPR-7694: Using environment variables to configure a jdbc url is now supported.
  • TPR-8330: Unexpected database exceptions during page-in operations are now handled more gracefully.
  • TPR-8997: Default behaviour of generated H2 databases changed.
  • TPR-9017: Cleaned up some unnecessary scripts in bin folder.
  • TPR-9074: Increased default connection timeout for SQLite paging databases. This prevents errors in case of long running save operations.
  • TPR-9244: Searching in BITARRAY Fields with minScore set to 100 now returns correct results.
  • TPR-9516: Uniqueness check for bean IDs is now case insensitive.
  • TPR-9517: It is now possible to restrict the space that databases occupy on the filesystem, depending on which database is used:
    • SQLite: The tool matchAdjustPaginDatabase can now be startet with the optional command –command VACUUM
    • H2: The SHUTDOWN command was extended to SHUTDOWN DEFRAG
    • PostgreSQL: VACUUM should be activated by default as long as not configured otherwise in the database.
  • TPR-9591: Fixed problem regarding Landscape Engine and usage with service name instead of SID.
  • TPR-9708: Performance test of Oracle databases via matchInspectPaginDatabase toll now returns all values correctly.
  • TPR-9721: Rule fields with zero weight are now ignored during the calculation of the total rule score with totalScoreMode=GEOMETRIC_MEAN.
  • TPR-9734: All non-empty match entries of a MULTISTRING field are now added if minScore is set to zero.
  • TPR-9842: Fixed problem with executing of the checkglibc.sh tool.
  • TPR-9973: Switched to UCP datasource for Oracle databases, since the old driver is deprecated.
  • TPR-10000: Rule-fields with minScore=0 are now ignored during the paritioning. This means, that all partions are considered for fields with minScore=0.

Known Bugs

  • TPR-10122: The wizard for executing background jobs via GUI does not scroll properly to the bottom in case of long console outputs.
  • TPR-10123: Long batch runs (30min +) via GUI can lead to lost sessions.

Archived release notes