It seems there are two packages providing access to TCGA data these days:
Will GenomicDataCommons supersede TCGAbiolinks? Is there anything that I might miss if I decide to use only GenomicDataCommons?
There isn't a R/Bioconductor person or view, so there isn't a preferred way to access TCGA. Use what suits you.
Bioconductor is build by contributors which are encourage to not overlap in their contributions. Usually the packages have different options/features even if they allow to access to the same data.
In this case, the TCGAbiolinks package is specifically for TCGA dataset while GenomicDataCommons is for "access the NIH / NCI Genomic Data Commons RESTful service". So the functionality they bring to R and Bioconductor is different.
The fact that one project is finishing doesn't mean that the data or the packages build around that project are obsolete. Some functionality different between them is the functions to harmonize and prepare (
GDCprepare) data from the TCGAbiolink package, or see a boxplot correlation and AAIC plot to define outliers