[DataFrame] what about integration of dataframe in Gama?

[chapuisk]

Matrix structure in Gama cannot be manipulated based on explicit headers, like it is usually the case in structure files such as csv, json and others. Matrices in Gaml are also limited in the number of linked data it can store, only 2 dimension. Hence implementing a DataFrame in Gaml could be a great addition manipulating complex data, as well as connecting with extensive ML libraries. This last point could be a game changer when it comes to simulation analysis or population synthesis in Gama (just to talk about the part I have been working on). Furthermore, I know that most of Gama users are familiar of language such as R or Python, even before they learn Gaml. Giving them the ability to manipulate known complex structure could be a plus.

The potential best option is Apache Spark data frames. What is your opinion about this? Should we go for a new gaml type or expends on already existing matrices?

Best
Kevin