LoginSign Up
← Back to all sparks
F

fmtr

ANALYTICS
Velocity0.0

Easily Apply Formats to Data

github.com/dbosak01/fmtr ↗

Rebuilding SAS's formatting layer in R, one format specification at a time

sas-paritydata-formattingclinical-reportingformat-cataloguesr-package
◆Current state
fmtr applies formats to R data the way SAS applies them: named format catalogues, format lists, and an fapply() that maps a specification onto a vector. It is part of a family of packages that reconstruct SAS reporting idioms in R, and it shares infrastructure with them — labels.data.frame() was moved out to the common package, which fmtr now depends on. The recent releases have been closing specific gaps against SAS's own format vocabulary.
◆Where it's heading
The direction is parity, pursued in small increments. Quarter format codes were added because base R has none; the SAS best. format was reimplemented, then hardened against the variations people actually write; statistical summary helpers like fmt_mean_sd() and fmt_mean_stderr() cover the cell contents clinical tables need. The structural work is largely behind it, including the breaking 2022 move that handed labelling to a sibling package, so what remains is vocabulary coverage.
◆Prediction
The pattern of adding a SAS format, then a release to handle its variants, suggests the next releases continue filling in format codes and summary helpers rather than changing how formats are applied.

◆Recent moves

  1. 10mo ago

    Mean and standard error helper, plus best-format variants

    Adds fmt_mean_stderr() alongside the existing mean and standard deviation helper, and handles the variations on the best specification that the previous release did not anticipate. The standard one-two of this package: add a SAS-derived format, then absorb how people actually write it.

    View source ↗
  2. 11mo ago

    SAS best. format reimplemented in fapply()

    Recreates SAS's best. format, which picks a sensible numeric representation without the caller specifying width or decimals. A direct parity addition of the kind that defines this package's roadmap.

    View source ↗
  3. 2y ago

    Quarter format codes %q and %Q added

    Introduces quarter format codes that base R simply does not provide, filling a gap that matters for any reporting on fiscal periods. Small in size, but it extends the format vocabulary rather than adjusting existing behaviour.

    View source ↗
  4. 2y ago

    Format lists become readable and writable files

    Adds read.flist() and write.flist() so format lists persist to disk, allows single-item vector lookups, and lets a Factor column drive the as.factor property when converting data frames. Persisting format catalogues is what makes them reusable across programs, which is how they work in SAS.

    View source ↗
  5. 2y ago

    Documentation and examples expanded

    Documentation and example improvements with no functional change. Notable only as the quiet gap between the 2023 and mid-2024 feature releases.

    View source ↗
  6. 2y ago

    value() can return results as a factor

    Adds an as.factor parameter to value() so formatted results come back as factors, which controls ordering downstream in tables and plots. The same property later became settable from a data frame column.

    View source ↗