Hi every one. Metaflow is a great product, and I succeed to start the dev stack. However, I did not undertand in doc how I can manage binary data. Here is my usecase : 1. I load a CSV in a df pandas. One column is a

filename

2. I have to load the

filename

and process him (calculate the len for instance) Than I created the following flow : 1. Loading DF -> iterating over the

filename

columns to the next step with

id

as input 2. Load the file using the

self.input

, compute the length, store it in

self.len

as the id in

self.id

3. Converge to the next step that must reconcile the length iterating on inputs

self.df.loc[input.id, 'len'] = self.len

However, the number of filesto process is huge, and I'm thinking to put the process in cloud. But the files are only on the host ... What's the best way to proceed ?

Can you put the files on s3?