Semantic search

Jump to navigation Jump to search

Significant Improvements on the Lookup Throughput (17 February 2021)

Upgrades and optimizations on our underlying indexing infrastructure allow us now for a much higher performance of our Business Partner Lookup. In batch mode >80 requests per second can now be handled easily. For any other CDQ Cloud Services that rely on connected data sources this will lead to much faster results (e.g. identifier or legal address enrichment). Please note that for data sources to which we connect directly via web interfaces such as VIES we are limited by the response times of the corresponding services. The improvements just have an impact on data sources which we manage in our own infrastructure.

Changes on the Duplicate and Linkage Matching App (16 February 2021)

The Duplicate and Linkage Matching App have been revised to follow the same design like e.g. the Batch Data Validation App (see https://meta.cdq.com/Release_note/20210128). These two apps are now structured by Storage Selection, Job Selection and Report Selection. Jobs can now be also started only for specific countries. In this context new duplicate and record linkage reports have been introduced. They follow the identical structures of the legacy reports but will allow for faster change management in the future. Moreover, the reports come along with charts and figures for gaining a quick overview on the results. The reports are also designed and prepared for upcoming releases on considering matching group reviews in subsequent duplicate identification jobs. Additionally, these reports can now be generated by users directly via our matching API: https://www.apimatic.io/apidocs/data-matching/v/2_0#/rest/api-endpoints/analytics/build-matching-reports

Selected fixes and improvements on Data Quality Rules (12 February 2021)

We improved some of our Data Quality Rules that were providing ambiguous results. The checkdigit check for the Indian GST Number was returning wrong results in seldom cases (https://meta.cdq.com/Identifier_checkdigit_invalid_(GST_number_(India))). The implementation of the underlying algorithm was fixed. Additionally, data quality rules that check if PO Box and care of information is misplaced in e.g. the business partner's name now make use of our Regular Expression Library (https://meta.cdq.com/PO_Box_information_misplaced_(Name_LOCAL)). In this context some fuzzy parsing cases were improved (e.g. BOX as part of a correct business partner name).