upcoming sf breaks ndhplusTools

[edzer]

with:

Package: nhdplusTools
Check: tests
New result: ERROR
    Running ‘testthat.R’ [71s/44s]
  Running the tests in ‘tests/testthat.R’ failed.
  Complete output:
    > library("testthat")
    > 
    > test_check("nhdplusTools")
    Loading required package: nhdplusTools
    Starting 2 test processes
    [ FAIL 1 | WARN 41 | SKIP 51 | PASS 296 ]
    
    ══ Skipped tests ═══════════════════════════════════════════════════════════════
    • On CRAN (51)
    
    ══ Failed tests ════════════════════════════════════════════════════════════════
    ── Failure ('test_index.R:195'): get location ──────────────────────────────────
    sf::st_coordinates(locations)[, 1:2] (`actual`) not equal to structure(...) (`expected`).
    
    `dimnames(actual)[[1]]` is NULL
    `dimnames(expected)[[1]]` is a character vector ('1', '2', '3')
    
    [ FAIL 1 | WARN 41 | SKIP 51 | PASS 296 ]
    Error: Test failures
    Execution halted

in NEWS.md:

• the matrix returned by st_coordinates() has no row names, to reduce output size

The row.names were meaningless anyway, and for large datasets take up a large amount of memory.

[dblodgett-usgs]

Thanks for the heads up -- when will this be hitting CRAN?

[edzer]

It's in the queue, so that might be really soon.

[dblodgett-usgs]

hrmmm... I'm having a hard time installing from github. It appears that this is just a test that is over specified though.

[edzer]

I'll check.

[dblodgett-usgs]

I think this should do it. dblodgett-usgs@7fd6be6#diff-739429cefec135bb4c884b1947eb14fbe0938e19021f62bbb589c7bd1a77c44dL195

[edzer]

Column names are still there, so the RHS of the test should be

structure(c(-76.8693957911233, -76.9176139910484, -76.8810037244386,
                           39.4932572053652, 39.4090934721626, 39.3632976055671),
                         .Dim = 3:2, .Dimnames = list(NULL, c("X", "Y")))

will check locally.

[dblodgett-usgs]

Yeah -- I just converted to numeric as my test is only really needing to checking that the numbers come back the same.

[dblodgett-usgs]

0.6.2 is on its way to CRAN. Thanks for the heads up @edzer