[R] Tests failures on R < 4.0 due to data.frame conversion

[thisisnic]

Describe the bug, including details regarding any error messages, version, and platform.

The test-r-versions nightly build fails due to a change in our tests introduced in #35173 which induces a test failure in older R versions.

══ Failed tests ════════════════════════════════════════════════════════════════
── Failure ('test-dataset-csv.R:94:3'): CSV scan options ───────────────────────
`object` (`actual`) not equal to `expected` (`expected`).

`actual$chr` is a character vector ('foo', NA)
`expected$chr` is an S3 object of class <factor>, an integer vector
Backtrace:
    ▆
 1. └─arrow:::expect_equal_data_frame(...) at test-dataset-csv.R:94:2
 2.   └─arrow:::expect_equal(as.data.frame(x), as.data.frame(y), ...) at tests/testthat/helper-expectation.R:24:2
 3.     └─testthat::expect_equal(...) at tests/testthat/helper-expectation.R:43:4

[ FAIL 1 | WARN 10 | SKIP 32 | PASS 7814 ]
Error: Test failures
Execution halted

Component(s)

R

[paleolimbot]

Somewhere we need a stringsAsFactors = FALSE!