A Generative AI Engineer has successfully chunked unstructured documents and wants to store the chunks in a Vector Search index. The current dataframe has columns for the original document name and an array of text chunks. What is the most performant way to store this dataframe?