Is it possible to get table name to be used in dbplyr::sql

[era127]

I am trying to add an operation to the end of my pipeline that uses the duckdb PIVOT expression.

To use dbplyr::sql() with a manual sql query I would need to get the name of the table that is used in the query which is created by as_duckplyr_df(). Is that possible?

Alternatively, is it possible to map the duckdb PIVOT expression to dplyr verbs?

[krlmlr]

Thanks, interesting use case. None of this seems to be available out of the box, but making a data frame available in duckdb via duckdb::duckdb_register() is instantaneous, perhaps with a small performance overhead during runtime.

The first option would be the equivalent of compute(), with a way to feed the result into an SQL expression. Let me think about an API here.

Ideally, we would support pivot_wider() and pivot_longer() natively, this seems to be much further down the road.