IntegrateData memory usage

[TaurVil]

Hello, the pipeline for data integration seems extremely memory intensive between the FindIntegrationAnchors and IntegrateData steps. Are there any ways these pipelines could be modified to run in a way that minimizes RAM? For example, if IntegrateData would save intermediate files, purge the working memory, and re-read in intermediate files when necessary later. I understand this would increase computation time, but it would also enable more flexibility and avoid issues such as a 3Gb dataset ballooning beyond 32Gb of RAM if there are a lot of files to be integrated.