Solved

Data in 'Learn to Use the Cognite Python SDK' is not as intended

Forum|Forum|2 years ago
April 10, 2024
6 replies
97 views

Hi, I’m doing the Learn to Use the Cognite Python SDK module in the Data Scientist basics seems incorrect.

For example, I’m in the notebook `2_List_Search_Retrieve.ipynb` and the data presented appears to be very random, and not industrial data. When I run a query like:

`c.assets.list(metadata={'ELC_STATUS_ID': '1211'},limit=5)` this comes back empty because there is no industrial data in any of the the datasets. In fact, almost all of the datsets in the c.data_sets.list() command are Deprecated and archived.

It looks to me like potentially the datasets are tests that people have made.

I would really like to be able to see some more realistic data in the dataset used for exploring the tool. Please let me know how I can do that, if anyone knows!

data science

Best answer by Sofie Haug

Thanks for the info @lchutny

As you follow along the notebooks, you will write some data yourself. This results in there being some additional “random” data in there generated by the learners. We do clean up every 14 days, but there will always be some extra data there.

I will need to review the notebook for the “empty” answers you mentioned. I hope you still got the chance to learn to use Cognite Python SDK.

Sofie Haug

Cognite Academy Instructor

Hi @lchutny

Are you using the notebook that belongs to the course? From the authentication part - could you tell me the cdf-project you are connected to?

Sofie | Cognite Academy

L

lchutnyAuthor

Hi, Sofie. Yes, I’m using the notebook as downloaded from the Github repo. The cdf project is `ds-basics`

CDF_CLUSTER = "westeurope-1"

COGNITE_PROJECT = "ds-basics"

BASE_URL = f"https://{CDF_CLUSTER}.cognitedata.com"

SCOPES = [f"{BASE_URL}/.default"]

LauraC

Sofie HaugAnswer

Cognite Academy Instructor

Thanks for the info @lchutny

As you follow along the notebooks, you will write some data yourself. This results in there being some additional “random” data in there generated by the learners. We do clean up every 14 days, but there will always be some extra data there.

I will need to review the notebook for the “empty” answers you mentioned. I hope you still got the chance to learn to use Cognite Python SDK.

Sofie | Cognite Academy

L

lchutnyAuthor

Hi, Sofie. Finally getting back to this, and one of the places that there are empty cells in workbook 2: metadata Filtering - essentially there are no metadata keys

c.assets.list(limit=5).to_pandas()['metadata'][4]

returns no results. So the subsequent list returns nothing either:

c.assets.list(metadata={'ELC_STATUS_ID': '1211'},limit=5)

But I’m still able to work through the notebook.

LauraC

Sofie Haug

Cognite Academy Instructor

Thanks for the heads up @lchutny

I’ll add it to my to-do list to get i checked. In the meantime just continue.

Sofie | Cognite Academy

L

lchutnyAuthor

@Sofie Haug a few more for you:

none of the datasets have any metadata, so the section on filtering using metadata returns nothing.
None of the timeseries have step-wise data, so the filter on ‘is_step=True’ returns nothing
The datasets from the Open Industrial Data project seem to be missing, a fuzzy search on name =”23-HA-9114” returns nothing.

Subsequently the ‘Try it yourself’ section where you have to search for a compressor and find all time series in the asset’s subtree doesn’t work. Asset “2ND STAGE COMPRESSOR-PH” doesn’t exist in the ds-basics workspace any longer.

LauraC

Sign up

Welcome to Cognite Hub