This is divided by type If we divide it from the operation after accessing the data source there are two categories those with table structure and those without table structure There is a table structure Those with table structures can be relational databases HIVE Doris etc that have table structures themselves It can also be fixedformat text or JSON that can be assigned a fixed schematic This type of data requires the data platform to have metadata management capabilities
This part will be introduced in When we talk about metadata Austria WhatsApp Number what are we talking about This type of table structure is interacted with in the wizard or drag and drop in the form of a twodimensional table No table structure Those without a table structure are relatively more complicated Sometimes you can force a table structure to be given to such a thing without a table structure Sometimes it can only be converted into a script to achieve mapping
detail How much the data source supports reflects the strength of the capability Similarly as a product each data source may have its own characteristics and require personalized design The product manager will be familiar with various types of data sources Personally I feel that this is also a part of the design of data integration products Troublesome point As for various unstructured documents pictures audio and video etc They are not within the scope of big data platform I have also mentioned unstructured big data platforms and unstructured big data governance before.