Sharpdata/npgsql error

Hello,
I have a PostgreSQL database that contains a number of tables, that contain columns of types such as geometry(Point),ltree and so on.
When I invoke SharpData with the command,

x open sharpdata -db postgres -db.connection $MYDB

On these tables, I see the error that is documented at:
https://www.npgsql.org/doc/faq.html#a-nameunknowntypei-get-an-exception-the-field-field1-has-a-type-currently-unknown-to-npgsql-oid-xxxxx-you-can-retrieve-it-as-a-string-by-marking-it-as-unknowna

http://localhost:5000/db/main/mytable?format=json&take=100
Error messages reads as follows:
The field ‘xxxx’ has type ‘public.geography’, which is currently unknown to Npgsql. You can retrieve it as a string by marking it as unknown, please see the FAQ.

I reviewed the code at:

Based on the code,
using args.tables, such as in the example,
args.tables Customer,Order,OrderDetail,Category,Product,Employee,EmployeeTerritory,Shipper,Supplier,Region,Territory
it is possible to filter tables shown in UI.

Is it possible to similarly, configure specific columns to be configured such that they are either completely ignored
or converted to string using a default/custom javascript function?

Thanks.

As Npgsql doesn’t support the type by default you would need to basically do what they specify in the docs, modify the query to cast the column as a text in sql, e.g. SELECT field::TEXT or use Npgsql APIs directly.

For Sharp Data the only way I can see to potentially avoid this issue with the generic _db/_table/index.html is to specify the list of fields, e.g. ?fields=Customer,Order,OrderDetail,... otherwise by default it falls back to select * all columns which will include the unsupported columns.

Otherwise you’d need to take modify a local copy of https://github.com/NetCoreApps/SharpData that somehow tries to support this specifically.