test.RdAn internal testing function used in data.table test scripts that are run by test.data.table.
test(num, x, y = TRUE,
error = NULL, warning = NULL, message = NULL,
output = NULL, notOutput = NULL, ignore.warning = NULL,
options = NULL, env = NULL)A unique identifier for a test, helpful in identifying the source of failure when testing is not working. Currently, we use a manually-incremented system with tests formatted as n.m, where essentially n indexes an issue and m indexes aspects of that issue. For the most part, your new PR should only have one value of n (scroll to the end of inst/tests/tests.Rraw to see the next available ID) and then index the tests within your PR by increasing m. Note – n.m is interpreted as a number, so 123.4 and 123.40 are actually the same – please 0-pad as appropriate. Test identifiers are checked to be in increasing order at runtime to prevent duplicates being possible.
An input expression to be evaluated.
Pre-defined value to compare to x, by default TRUE.
When you are testing behaviour of code that you expect to fail with an error, supply the expected error message to this argument. It is interpreted as a regular expression, so you can be abbreviated, but try to include the key portion of the error so as not to accidentally include a different error message.
Same as error, in the case that you expect your code to issue a warning. Note that since the code evaluates successfully, you should still supply y.
Same as warning but expects message exception.
If you are testing the printing/console output behaviour; e.g. with verbose=TRUE or options(datatable.verbose=TRUE). Again, regex-compatible and case sensitive.
Or if you are testing that a feature does not print particular console output. Case insensitive (unlike output) so that the test does not incorrectly pass just because the string is not found due to case.
A single character string. Any warnings emitted by x that contain this string are dropped. Remaining warnings are compared to the expected warning as normal.
A named list of options to set for the duration of the test. Any code evaluated during this call to `test()` (usually, `x`, or maybe `y`) will run with the named options set, and the original options will be restored on return. This is a named list since different options can have different types in general, but in typical usage, only one option is set at a time, in which case a named vector is also accepted.
A named list of environment variables to set for the duration of the test, much like options. A list entry set to NULL will unset (i.e., Sys.unsetenv) the corresponding variable.
NA_real_ and NaN are treated as equal, use identical if distinction is needed. See examples below.
If warning= is not supplied then you are automatically asserting no warning is expected; the test will fail if any warning does occur. Similarly for message=.
Multiple warnings are supported; supply a vector of strings to warning=. If x does not produce the correct number of warnings in the correct order, the test will fail.
Strings passed to notOutput= should be minimal; e.g. pick out single words from the output that you desire to check does not occur. The reason being so that the test does not incorrectly pass just because the output has slightly changed. For example notOutput="revised" is better than notOutput="revised flag to true". notOutput= is automatically case insensitive for this reason.
Logical TRUE when test passes, FALSE when test fails. Invisibly.
test = data.table:::test
test(1, x = sum(1:5), y = 15L)
test(2, log(-1), NaN, warning="NaNs")
test(3, sum("a"), error="invalid.*character")
# test failure example
stopifnot(
test(4, TRUE, FALSE) == FALSE
)
#> Test 4 ran without errors but failed check that x equals y:
#> > x = TRUE
#> First 1 of 1 (type 'logical'):
#> [1] TRUE
#> > y = FALSE
#> First 1 of 1 (type 'logical'):
#> [1] FALSE
#> 1 element mismatch
# NA_real_ vs NaN
test(5.01, NA_real_, NaN)
test(5.03, all.equal(NaN, NA_real_))
test(5.02, identical(NaN, NA_real_), FALSE)