Troubleshooting access on BigQuery

[rajsinghUSA]

In following the instructions here, after clicking either of the the "+ ADD DATA" buttons shown here:

After clicking the "+ ADD DATA" button, a panel is displayed (shown in the following screenshot).
However, you can see there is no "Pin a project" button/option

Am I missing some step? Or has the flow in BigQuery changed?

[briangow]

Yes, the BigQuery interface has changed. Please select "Search for and star a project" from your second screen shot. This will do the same thing as "pin" previously did.

[rajsinghUSA]

Thanks @briangow
I still do not see any results for "physionet-data" to Star:

[briangow]

Have you already completed the training and credentialing requirements and followed all of the instructions outlined here: https://mimic.mit.edu/docs/gettingstarted/cloud/ ?

[rajsinghUSA]

Yes, training and credentialing has been completed. In fact, I was able to access the MIMIC data on AWS, just not on BigQuery.

You can also see in this screenshot that the "Files" section is not red:

[briangow]

Did you also follow these instructions:https://mimic.mit.edu/docs/gettingstarted/cloud/ ? Do you have a gmail on your PhysioNet account: https://physionet.org/settings/emails/? What happens when you click "Request access using Google BigQuery" in your last screenshot?

[rajsinghUSA]

Yes, I already have the gmail linked in my profile.

Clicking on "Request access using Google BigQuery" shows this:

[rajsinghUSA]

In the PhysioNet GCP BigQuery access email, it says:
"If you cannot add the project "physionet-data", please access it through this link: https://console.cloud.google.com/bigquery?project=physionet-data&page=dataset".

Going to that link, I see some MIMIC3 data available:

[briangow]

Hmm, I'm not sure why you can't see physionet-data then. I assume you are signing into http://console.cloud.google.com/bigquery with the same gmail that was awarded access by clicking "Request access using Google BigQuery".

[rajsinghUSA]

Yes, same google account. And you can see it's letting me have some access to MIMIC3 via https://console.cloud.google.com/bigquery?project=physionet-data&page=dataset

[briangow]

@rajsinghUSA , my apologies, after speaking with my colleague it appears that we missed a step with the cloud setup when releasing v2.1 of MIMIC. Could you please try to +ADD DATA and search for physionet-data again.

[rajsinghUSA]

Great! I see there is now mimiciv available when going to https://console.cloud.google.com/bigquery?project=physionet-data&page=dataset :

And if choosing Star a project by name, then "physionet-data" can be entered.

Thank you!

[briangow]

Great, thanks for your patience and feedback!

[mjbroerman]

Hi! I think the maintainers may have a similar situation for the "derived" datasets. Also, I used the method above to find physionet-data and star it, but the search function did not work for it or the derived datasets.

I tried this url:

https://console.cloud.google.com/bigquery?project=physionet-data.mimiciii_derived&page=dataset

And it displays

However, you cannot star it as the resource does not seem to exist.

Am I correct that accessing this derived resource is the best way to run the queries for and other severity scores?

Thanks for maintain this wonderful resource!

[alistairewj]

Yes you're correct. You need to separately be provided access to MIMIC-III in order to view the mimiciii_derived table however. Can you confirm you've been approved for access to MIMIC-III, successfully requested access to MIMIC-III on BigQuery, but still can't see the mimiciii_derived dataset?

[RanElgedawy]

I am trying to access the MIMIC-IV-note data from a jupyter notebook. I keep getting a 403 error: 403 Access Denied: Table physionet-data:mimiciv_note.discharge: User does not have permission to query table physionet-data:mimiciv_note.discharge, or perhaps it does not exist in location US.

Would some one please tell me what I might be doing wrong here? @alistairewj @briangow

[tompollard]

The process for pinning a project on BigQuery has recently changed. See: #1453 for a step-by-step description of the new process.

[alistairewj]

It's definitely going to be an issue with the credentials. It's more typical to do gcloud auth with your Google account and let the default credentials be imported by python. Might be worth trying that and not manually setting credentials with a JSON file.