And I feel that this is too extreme
Although data synchronization is often compared with **ETL extract transform load** but in fact it is It is debatable that conversion is not required during synchronization One to one synchronization The purpose of synchronizing data is to preserve the data history of the business If you want to preserve the history then the history of errors is also history Therefore this kind of synchronization is completely onetoone synchronization with the business system data even if the synchronizeddata is abnormal or nonstandard Only in this way can the Austria WhatsApp Number business history be truly preserved and when data anomalies occur and data is traced the original business data can be found Personally I think this is a good idea as it can completely retain business historical data But there is a problem wrong data business systems can be changed at any time However synchronization in offline scenarios will not be performed at any time and the requirements for personnel and procedures are relatively high
https://lh7-us.googleusercontent.com/4FckHUwXk3b92snLoDsNzrVJcXT1eSqDqWXgeJIB4sT-A1jo3qhBhnNIWbIBnkPMRRlWQ5LQRsfWAYdlUn6VTcUfU72jwfONZg5faKDa3GIYAxGBJPDHtuDWwZ0FYxr_cyqhb24I_k9lRKAzhefNi1Y
Conversion cleaning during synchronization The second kind of requirements are obviously not so strict and are relatively loose Rowlevel additions subtractions and normalization can be performed during this process You can also perform operations such as field aggregation association conversion etc In fact for product design if this form is supported onetoone synchronization is supported There is this conversion and aggregation capability during the synchronization process If not used it will be onetoone synchronization In this way onetoone synchronization seems to be more of a norm and a requirement .
頁:
[1]