Need improvement on creating new feature in Feature Store documentation

I am talking about Feature Store feature, which is document in here

In “Create new feature” section, the documentation missing the part how to load dataframe from file into notebook


More specific, in the screenshot (version 1.4), how to get “features_df” when all you have is the .csv file in your PC.


My step-by-step

  1. Go to Data Sets
  2. Upload iris.csv
  3. Start jupyter
  4. (stuck) trying to load iris.csv from Data Set but don’t know how to ? (what is the code to read, what is the full path to iris.csv)

Here, Data Sets in file description I also suggest adding a snipping of code on how to load to jupyter, or at least a path

Hi Thaithien,

copy the link and read csv by

val df = spark.read.option("header",true)
   .csv("path_to_csv")

then

Hops.createFeaturegroup("your_csv_fg").setDataframe(df).write()

Let me know if this works.

You can also find more examples in this notebook.

Cheers,
Davit