In-memory vs in-db methods

[Matt Seminaro]

From my understanding, databases can be loaded using an in-db or in-memory connection. Since databases can be loaded with an in-memory connection, what is the diffrence between an in-memory database connection, an information link, and loading data through data functions. Addtitionally,I have not seen much information on loading data throughdata functions, what are the pros and cons of this method compared to the other two

[Danny Yu]

When loading data by the database connector, the data actually loaded directly from source to the Client memory, and with an option of in-DB and in-memory, for the possible big data use cases and analytics reporting.

On the other hand, the information link is to load the data from the source through the server, so that it can be shared between authorized users and different dxp, but only in-memory mode can be used. Mostly it will be used when the dxp is a kind of operational/regular reporting.

For loading data via data function (by Python or R), it would be used if the data needs to be pre-processed and/or aggregated by a certain algorithm, or no suitable connector can be used, so that you need to develop your own.

I think those 3 ways are all different purposes, pros and cons are not the key points to make a choice. It all depends on your use case.

 

-Danny Yu