Seasoned ⭐️

Solved

PI Extractor - Backfilling issues

Forum|Forum|13 days ago
August 11, 2026
9 replies
131 views

Hello!

We have encountered some issues with backfilling on the PI Extractor. This usually goes automatically, and works fine. However, we do now have a 2 day gap in data on Cognite, however all that data is present in PI.

Are there any way to “force” a backfill, or fix this issue somehow?

Thanks!

Best answer by Mithila Jayalath

@gorankarlsen can you try out Cognite PI Replace Utility. It allows you to target a specific date range, optionally delete the "straight-line" gap in CDF, and re-ingest the actual historical data from your PI Server.

Here is a configuration template:

version: 2
cognite:
  project: <your-project>
  idp-authentication:
    tenant: ${COGNITE_TENANT_ID}
    client-id: ${COGNITE_CLIENT_ID}
    secret: ${COGNITE_CLIENT_SECRET}
    scopes:
      - ${COGNITE_SCOPE}
time-series:
  external-id-prefix: "pi:"   # Ensure this matches your PI extractor's prefix
pi:
  host: ${PI_HOST}
  username: ${PI_USER}
  password: ${PI_PASS}
replace:
  start-time: "<gap-start-time>"
  end-time: "<gap-end-time>"
  delete-data-points: true   # Deletes the straight-line gap in CDF before re-ingestion
  include-tags:
    - "<YOUR_TAG_1>"
    - "<YOUR_TAG_2>"

Recommended steps:

Dry-run first: Add dry-run: true to the replace block and run PiExtractor.exe replace -f <your_config_file>.yml. This will compare PI data with CDF data and log differences without making any changes.
Execute: Once you've verified the dry-run output, set dry-run: false (or remove the line) and run the command again to perform the actual replacement.

Mithila JayalathAnswer

Expert ⭐️⭐️⭐️⭐️

@gorankarlsen can you try out Cognite PI Replace Utility. It allows you to target a specific date range, optionally delete the "straight-line" gap in CDF, and re-ingest the actual historical data from your PI Server.

Here is a configuration template:

version: 2
cognite:
  project: <your-project>
  idp-authentication:
    tenant: ${COGNITE_TENANT_ID}
    client-id: ${COGNITE_CLIENT_ID}
    secret: ${COGNITE_CLIENT_SECRET}
    scopes:
      - ${COGNITE_SCOPE}
time-series:
  external-id-prefix: "pi:"   # Ensure this matches your PI extractor's prefix
pi:
  host: ${PI_HOST}
  username: ${PI_USER}
  password: ${PI_PASS}
replace:
  start-time: "<gap-start-time>"
  end-time: "<gap-end-time>"
  delete-data-points: true   # Deletes the straight-line gap in CDF before re-ingestion
  include-tags:
    - "<YOUR_TAG_1>"
    - "<YOUR_TAG_2>"

Recommended steps:

Dry-run first: Add dry-run: true to the replace block and run PiExtractor.exe replace -f <your_config_file>.yml. This will compare PI data with CDF data and log differences without making any changes.
Execute: Once you've verified the dry-run output, set dry-run: false (or remove the line) and run the command again to perform the actual replacement.

G

gorankarlsenAuthor

Seasoned ⭐️

@Mithila Jayalath Hello! Thank you for posing this. I have now tried replace, but receiving some sort of error. Also, seems like this does not work for config version 3. Any other workarounds for this matter?

Yaseen Ahmed Khan

Practitioner ⭐️⭐️⭐️

Hi @gorankarlsen,
PI Replace accepts version: 2 , and the main extractor takes the version: 3
So, changing the version to 2 as mentioned in @Mithila Jayalath’s configuration template should resolve that error.
Please let us know if it worked once you test it out.
Thanks!

G

gorankarlsenAuthor

Seasoned ⭐️

Hello! @Yaseen Ahmed Khan @Mithila Jayalath

Thank you for the support. I tried editing to version 2. I now did not receive any error in the command prompt; however, no data was backfilled in cognite. Not sure where this goes wrong for me; as this should automatically backfill by default?

Yaseen Ahmed Khan

Practitioner ⭐️⭐️⭐️

Hello @gorankarlsen
Could you confirm if the dry-run is set to false and the start-time/end-time matches your actual gap period. Could you also share the configuration along with the log ?
Thanks

G

gorankarlsenAuthor

Seasoned ⭐️

Hello! @Yaseen Ahmed Khan I can share the logs etc, no problem. Just to clarify, I tried creating a new config for the replace method. Is that OK or do I have to change the original config file when doing the replace method? Then I used the -f to run the config created for this particular task.

Yaseen Ahmed Khan

Practitioner ⭐️⭐️⭐️

Hello @gorankarlsen, yes that should be fine, once your run is confirmed, if it still fails, then we would need to look at
1. config file
2. Exact command that was run
3. log of the run (terminal output or log file)

G

gorankarlsenAuthor

Seasoned ⭐️

@Mithila Jayalath @Yaseen Ahmed Khan Thank you both for your time and explanations. I eventually managed to configure the replace-config properly, and we were able to replace/backfill most signals. However, we received an error on a few pipoints, which complains it is unknown, allthough it is currently in both PI server and cognite.

Why that occurs, I am not sure, but we atleast got most data points backfilled, which is a good start.

Once again, highly appreciate your time and support.

Mithila Jayalath

Expert ⭐️⭐️⭐️⭐️

@gorankarlsen happy to hear that your issue is mostly resolved. If you come across any other issues, feel free to create a topic.

Sign up

Welcome to Cognite Hub