pivot_wider_spec() should check if data is a data frame

[mgirlich]

The error message when passing in a lazy_frame() or memdb_frame() are not very helpful:

spec <- tibble::tibble(
  .name = c("x", "y", "z"),
  .value = "val",
  key = c("x", "y", "z")
)

dbplyr::memdb_frame(key = c("x", "y", "z"), val = 1:3) |>
  tidyr::pivot_wider_spec(spec)
#> Error in `data[vec_unique(c(id_cols, names_from_cols, values_from_cols, unused_cols))]`:
#> ! Can't subset columns that don't exist.
#> ✖ Columns `src` and `lazy_query` don't exist.
#> Backtrace:
#>      ▆
#>   1. ├─tidyr::pivot_wider_spec(...)
#>   2. │ ├─...[]
#>   3. │ └─tibble:::`[.tbl_df`(...)
#>   4. │   └─tibble:::vectbl_as_col_location(...)
#>   5. │     ├─tibble:::subclass_col_index_errors(...)
#>   6. │     │ └─base::withCallingHandlers(...)
#>   7. │     └─vctrs::vec_as_location(j, n, names, missing = "error", call = call)
#>   8. └─vctrs (local) `<fn>`()
#>   9.   └─vctrs:::stop_subscript_oob(...)
#>  10.     └─vctrs:::stop_subscript(...)
#>  11.       └─rlang::abort(...)
dbplyr::lazy_frame(key = c("x", "y", "z"), val = 1:3) |>
  tidyr::pivot_wider_spec(spec)
#> Error in `as.data.frame()`:
#> ! Can not coerce <tbl_lazy> to <data.frame>
#> Backtrace:
#>     ▆
#>  1. └─tidyr::pivot_wider_spec(...)
#>  2.   ├─tibble::as_tibble(data)
#>  3.   └─tibble:::as_tibble.default(data)
#>  4.     ├─tibble::as_tibble(...)
#>  5.     ├─base::as.data.frame(value, stringsAsFactors = FALSE)
#>  6.     └─dbplyr:::as.data.frame.tbl_lazy(value, stringsAsFactors = FALSE)
#>  7.       └─cli::cli_abort("Can not coerce {.cls tbl_lazy} to {.cls data.frame}")
#>  8.         └─rlang::abort(...)

^{Created on 2023-07-13 with reprex v2.0.2}