strings_as_factors

Added in 0.6.0

What it does

Checks for calls to data.frame() that contain a statically identifiable character column but do not explicitly set stringsAsFactors. This rule only applies when the project’s minimum supported R version is known and is below R 4.0.0.

Why is this bad?

Before R 4.0.0, data.frame() converted strings to factors by default. From R 4.0.0 onward, strings remain character vectors by default. Code supporting versions on both sides of this change can therefore return columns with different types depending on the R version used.

This rule does not provide an automatic fix because either TRUE or FALSE can be the intended value of stringsAsFactors.

Example

data.frame(x = "a")

Use one of the following instead:

data.frame(x = "a", stringsAsFactors = TRUE)
data.frame(x = "a", stringsAsFactors = FALSE)

References

See ?data.frame and the R Core discussion.