FFlist(): scrapes Kenneth French’s data library and
returns a tidy data frame (or tibble) of all available datasets, with
columns name and file_url. The
name column can be passed directly to
inputlist.
FFmatch(): previews fuzzy-match results for a given
inputlist before downloading. Returns
requested, matched,
edit_distance, and similarity columns so users
can verify matches without committing to a download.
FFget(): downloads a single dataset directly into
the R session (no file I/O). Accepts frequency and
subtable arguments to slice the desired sub-table. Applies
na_values = c(-99, -999, -99.99) by default.
FFdownload()New na_values parameter: replace French’s sentinel
missing-value codes (e.g. −99, −999, −99.99) with NA after
numeric conversion.
New return_data parameter: when TRUE,
returns the FFdata list invisibly in addition to saving the
.RData file.
New action parameter: cleaner alternative to the
download / download_only flag pair. Accepts
"all", "list_only",
"download_only", or "process_only".
New cache_days parameter: skips re-downloading zip
files in tempd that are younger than the specified number
of days.
New match_threshold parameter: emits a warning when
the fuzzy-match similarity for any requested dataset falls below the
threshold.
converter() and converter_tbl() now accept
a na_values argument and apply sentinel replacement after
numeric conversion.assetpricing.Rmd: a complete empirical asset
pricing workflow covering time-series tests (CAPM, FF3) on 25 Size×BM
portfolios, the GRS joint test, the momentum anomaly, and Fama-MacBeth
cross-sectional regressions under CAPM, FF3, FF4 (Carhart), and FF5
across 100 Size×BM portfolios.donttest{} but because
cran tests this anyway (see
https://stackoverflow.com/questions/63693563/issues-in-r-package-after-cran-asked-to-replace-dontrun-by-donttest)
I also commented the second example which takes some time out (as
dontrun() seems to also not being liked on cran)NEWS.md file to track changes to the
package.