The hardware and bandwidth for this mirror is donated by METANET, the Webhosting and Full Service-Cloud Provider.
If you wish to report a bug, or if you are interested in having us mirror your free-software or open-source project, please feel free to contact us at mirror[@]metanet.ch.

tidyna

Tired of littering your code with na.rm = TRUE?

tidyna masks common R functions and warns you when NAs are removed. It handles some special cases. The table() default is set to useNA = "ifany".

Installation

Install from CRAN:

install.packages("tidyna")

Or install the development version from GitHub:

# install.packages("pak")
pak::pak("statzhero/tidyna")

Usage

library(tidyna)

x <- c(1, 2, NA)
mean(x)
#> ⚠️ 1 missing value removed.
#> [1] 1.5

Suppress warnings with options(tidyna.warn = FALSE).

Functions

• Summary: mean, sum, prod, sd, var, median, quantile
• Extrema: min, max, pmin, pmax, range
• Logical: any, all
• Row-wise: rowSums, rowMeans
• Correlation: cor
• Table: table

Special cases

All-NA input is configurable: By default, tidyna throws an error when all values are NA to prevent misleading values like Inf, NaN, or 0:

base::sum(c(NA, NA), na.rm = TRUE)
#> [1] 0

sum(c(NA, NA))
#> Error in `sum()`:
#> ! All values are NA; check if something went wrong.

You can change this behavior with the all_na argument or the tidyna.all_na option:

# Return base R behavior (NaN, Inf, 0, etc.)
sum(c(NA, NA), all_na = "base")
#> [1] 0

# Always return NA
sum(c(NA, NA), all_na = "na")
#> [1] NA

rowSums/rowMeans return NA for all-NA rows, but error if the entire matrix is NA. Also configurable via all_na.

pmax/pmin return NA for positions where all inputs are NA (with a warning), but error if every position is all-NA. Also configurable via all_na.

cor defaults to use = "pairwise.complete.obs" instead of erroring on NAs.

table defaults to useNA = "ifany", showing NA counts when present rather than silently dropping them.

Performance

There is no free lunch. The tidyna package adds some overhead:

For most functions like mean() the overhead is negligible (1.1x). But rowMeans() and rowSums() require an extra pass to detect all-NA rows, so there is a substantial loss (3-4x).

I’m still working on whether the memory allocation needs to be addressed.

Roadmap

• Add explicit _aware suffixed versions (mean_aware, sum_aware, etc.) for users who prefer not to mask base functions.

Related packages

• naflex: Conditional NA removal based on thresholds
• na.tools: Utilities for working with missing values

These binaries (installable software) and packages are in development.
They may not be fully stable and should be used with caution. We make no claims about them.