dataset-io

This is a (minimal) set of dependencies needed to enable tech.ml.dataset and tablecloth to use these file formats out of the box:

• Excel
• Arrow
• Parquet

How to use

Add this to deps.edn (apart from tech.ml.dataset or tablecloth):

scicloj/datase-io {:git/url "https://github.com/scicloj/dataset-io"
                   :git/sha "752bc3125c9701496a7ac0a2e722254bc349d2e9"}

The next call register handlers, so that ->dataset (TMD) or dataset (tablecloth) can read the file types above.

((requiring-resolve 'scicloj.dataset-io/enable-all-input-formats!))

Excel

(tablecloth.api/dataset "data/singleSheet.xlsx")

Load mulitple sheets via fastexcel:

(tech.v3.libs.fastexcel/workbook->datasets "data/twoSheets.xlsx")

Arrow

(tablecloth.api/dataset "/workspaces/tablecloth/data/alldtypes.arrow-feather"
                        {:file-type:arrow})

Parquet

(tablecloth/dataset "/workspaces/tablecloth/data/userdata1.parquet")