Accessing file assets in Spaces

When creating assets on the platform, you can create either file or table asset.

Table assets are those that have a recognised and supported format. Their schema and structure can be identified and they can be queried from structured databases where they’re stored. Table assets include data stored as csv, parquet, avro, orc or BigQuery or Snowflake assets.

File assets on the other hand can include any data type or format and are simply recognised and stored as objects in object storage.

Where are file assets stored?

File assets are stored as files under the following directory: /mnt/spaces/assets

You can list all of the assets by using a simple ls command in Jupyter or RStudio

There may be a single file, multiple files or multiple folders and files under the root directory of each product / asset that's added into the Space - depending of the make-up of the asset defined by the producer.

You can list this folder to see all of the file assets in your Space, from inside a Jupyter notebook, or using the Jupyter terminal.

Jupyter notebook cell: !ls /mnt/spaces/assets
Jupyter terminal: ls /mnt/spaces/assets
RStudio terminal: ls /home/rstudio/spaces_persistent_home/

Copying/loading the data into Jupyter or RStudio for use

Jupyter

In the example below there is one asset in the Space files_asset , which contains one JSON file - file.json

Using the full path, you can copy this into your working directory so that it appears in the Jupyter file browser: !cp /mnt/spaces/assets/files_asset/file.json /home/jovyan/workfile.json

Depending on the type of the file, you may want load it into a dataframe.

For example loading the json file into a pandas dataframe:

CODE

import pandas as pd

df = pd.read_json("/home/jovyan/work/file.json")

Note that you can also load the file directly from source without copying it out, so whenever the asset updates you are pulling the most recent version of the file.

CODE

import pandas as pd

df = pd.read_json("/mnt/spaces/assets/files_asset/file.json")

RStudio

The RStudio terminal can run through the same as above, you just need to point to a different working path /home/rstudio/spaces_persistent_home/

This will then appear in Jupyter as well as they have a shared working directory

For example in R terminal:
cp /mnt/spaces/assets/files_asset/file.json /home/rstudio/spaces_persistent_home/

You can also directly load the file into an R dataframe.

This example is JSON specific

CODE

install.packages("rjson")
library(rjson)

rDF <- fromJSON(file = '/mnt/spaces/assets/files_asset/file.json')