CODEWITHSUNDEEP
pythonjsonjupyter-notebookazure-blob-storageazure-machine-learning-service
anitasp
anitasp

Reputation: 419

Save a dict to json in Azure Machine Learning Jupyter Notebook

I built a dictionary with Jupyter Notebook in Azure Machine Learning Studio:

w_att = {
    '398465': 0,
    '8837.58': 1,
    '74967': 2,
    'jjpereza1': 3,
    '3180311358': 4,
    '56450': 5,
    '812723.990000033': 6,
    'guaba': 7}

w_att length is 1372600 so when I tried to store the object in the Notebook's instance with this code:

import json
json_object = json.dumps(w_att, indent = 4)
print(json_object)

I obtained this error:

> IOPub data rate exceeded.
> The notebook server will temporarily stop sending output
> to the client in order to avoid crashing it.
> To change this limit, set the config variable
> NotebookApp iopub_data_rate_limit.

> Current values:
> NotebookApp iopub_data_rate_limit 1000000 bytes sec
> NotebookApp rate_limit_window 3.0 secs

Then I tried:

import azureml.core
from azureml.core import Workspace, Datastore
import json

ws = Workspace.from_config()
datastore = Datastore.get(ws, datastore_name='xxx')
datastore.upload_files(json.dumps(w_att, indent = 4), overwrite = True)

And I got this error:

UserErrorException: UserErrorException:
    Message: '{' does not point to a file. Please upload the file to cloud first if running in a cloud notebook.
    InnerException None
    ErrorResponse 
{
    "error": {
        "code": "UserError",
        "message": "'{' does not point to a file. Please upload the file to cloud first if running in a cloud notebook."
    }
}

How can I directly save the object w_att to my storage account as a json file?

Upvotes: 1

Views: 986

Answers (1)

Anders Swanson
Anders Swanson

Reputation: 3961

Your dictionary is huge -- it looks like the notebook server fell down trying to return to you all the data.

To save out the file to your blob, you're almost there. You first have to save to a json object as an actual file then upload the file to the datastore. Check out this question on saving JSON to file in Python (turns out its more complicated than you would think).

import azureml.core
from azureml.core import Workspace, Datastore
import json

ws = Workspace.from_config()
datastore = Datastore.get(ws, datastore_name='xxx')

with open('data.json', 'w') as f:
    json.dump(w_att, f, indent = 4)

datastore.upload_files('data.json', overwrite = True)

Upvotes: 2

Related Questions