PLEASE HELP: Mongosqld.exe sometimes does not find data in collection/table

Wanted to get feedback from the community and support engineers as to what conditions could cause this anomaly to happen:

Mongosqld.exe finds all 21 of our collections/tables that have data and creates the relational schemas correctly.
There is however one table that sometimes it thinks does not have data.
This is a collection/table and not an object.
Apiedb.applications does in fact have data.
There are no changes being made to configurations to the connector or ODBC. No other deltas.

Mongosqld.exe sometimes thinks that a table DOES NOT have data:
2022-02-17T08:00:23.311-0800 D SCHEMA [sampler] mapping schema for namespace “apiedb”.“applications”
2022-02-17T08:00:23.407-0800 D SCHEMA [sampler] skipping namespace “apiedb”.“applications”: no documents found

Sometimes thinks that the table DOES have data:
2022-02-24T09:37:27.540-0800 D SCHEMA [mapping] mapped new table “applications”
2022-02-24T09:37:27.540-0800 D SCHEMA [sampler] finished mapping schema for namespace “apiedb”.“applications”

Any thoughts feedback greatly appreciated!

Additional note we are using MongoDB hosted in Azure.

@Mark_habana Hello and thanks for your post. My name is Alexi and I am the Product Manager for the BI Connector.

Based on what you have said, I am going to assume that you are using the self-managed BIC? And I believe you are saying that sometimes the schema shows this collection/table called apiedb.applications with docs/data and other times (without any overt reason) it thinks there are no documents.
You have stated this is in fact a collection, not a view, correct?

So I have a few initial thoughts - When the data is being resampled, it for some reason isn’t finding docs in this collection/table. How often are you sampling? “refreshIntervalSecs” and what does your sample size look like? There are some other runtime-options that you can tweak.

Are you using a customized DRDL file?

If not, this might help to control what the Schema looks like and what/when is sampled.

Is the data within this collection odd in anyway? Is it sparse, heavily nested etc.? The answer to this may help us to plug in options with the DRDL to make sure that data gets found anytime it is being sampled.

Let me know if any of this works. We also have a great support team that is well versed in this area.

Best of luck!

1 Like