Any idea how hard it would be to update WsSQL logic to allow queries against files with child datasets?

I believe my client would care less about being able to filter against child dataset rows/fields - but would probably be happy with my option 1, 2, or 3 above.

(edit...)

Rough analysis of the code that disallows child datasets in WsSQL...

When a query is sent to WsSQL for a file with child datasets, the esp.log shows this error message:

Nested data files not supported

https://github.com/hpcc-systems/HPCC-Platform/blob/c4487f7e71a37df7db20b3f3a6f0d985f6e5f2fc/esp/services/ws_sql/SQL2ECL/HPCCFileCache.cpp

...where the "containsNestedColumns" value is originally set here:

https://github.com/hpcc-systems/HPCC-Platform/blob/c4487f7e71a37df7db20b3f3a6f0d985f6e5f2fc/esp/services/ws_sql/SQL2ECL/HPCCFile.cpp

reopen if more info is needed.

 The Spark HPCC Connector does allow for reading / writing datasets with child datasets. This will automatically be handled and converted into the appropriate types for the Spark Dataframe.

WsSQL appears to NOT support child datasets in a select.

I.e., it fails doing a SELECT (even one that does NOT explicitly reference a child dataset field) against a logical file (table) that has child datasets.

Ways I can imagine to possibly "partially support" child datasets in WsSQL or even Spark HPCC Connector - allow SELECTs against these tables, but:

1. Include child dataset fields as Xml text in a single string field.

2. Include an empty field where the child dataset is.  (With a warning in the logs?)

3. Exclude the child data field.  (With a warning in the logs?)