ARROW-5190 [R]: Discussion: tibble dependency in R package

[romainfrancois]

tibble is now on Suggests. The code still makes tibbles, so that if tibble is otherwise loaded, data frames are nicely print, etc... but does not need to specifically Imports it.

[codecov-io]

Codecov Report

Merging #4454 into master will decrease coverage by 13.12%.
The diff coverage is 100%.

@@ Coverage Diff @@## master #4454 +/- ##
===========================================
- Coverage 88.11% 74.99% -13.13% 
===========================================
Files 850 54 -796 Lines 105774 3111 -102663 Branches 1253 0 -1253 ===========================================
- Hits 93203 2333 -90870 + Misses 12326 778 -11548 + Partials 245 0 -245

Impacted Files	Coverage Δ
r/R/Table.R	86.66% <ø> (ø)	⬆️
r/R/RecordBatch.R	96% <ø> (ø)	⬆️
r/R/feather.R	58.33% <100%> (ø)	⬆️
r/R/parquet.R	100% <100%> (ø)	⬆️
r/R/read_table.R	100% <100%> (ø)	⬆️
python/pyarrow/ipc.pxi
cpp/src/arrow/csv/chunker-test.cc
cpp/src/parquet/column_page.h
cpp/src/parquet/bloom_filter-test.cc
cpp/src/arrow/array/builder_decimal.cc
... and 791 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 48ee38f...18b1e05. Read the comment docs.

[nealrichardson]

LGTM; we aren't using tibble for anything special ourselves, just returning a data.frame with the tbl_df class, so it makes sense not to depend on it (not that the weight of the dependency is so significant, relative to libarrow :)

This does make the as_tibble arguments in read_feather/parquet/etc. slightly misnamed, but I'm not too concerned because I'm hoping we can remove that arg soon. IMO read_feather et al. should always return a data.frame, and we should provide a different function (I've proposed open_dataset()) for opening a file or directory of files (local or remote) in Arrow-space.