Hi Carlo,
One way to solve this issue could be by splitting the original parquet file into multiple smaller files and create “parquetDatastore” of them. The data can then be obtained from the datastore using the “read” function. A datastore is a repository of files where each individual file can fit in the memory but the entire collection of files may not fit. They are used to work with the data incrementally and hence reduces the memory usage.
You can refer to the following link for more information on datastores: https://www.mathworks.com/help/releases/R2022b/matlab/ref/datastore.html