ARROW-3942: [R] Feather api fixes

[romainfrancois]

Some fixes to follow up open #3043, and added the columns argument to read_feather that can be:

• character vector
• integer vector : 1-based in R
• NULL: to get all columns (the default)

Also adds as_tibble argument to read_feather to switch between data.frame and arrow::Table return value

[wesm]

I had suggested having an as_tibble = TRUE (or as_arrow = FALSE) argument in read_feather to preserve the default behavior of feather::read_feather. Thoughts?

[romainfrancois]

Then I suppose it would also need use_threads. This only apply to the conversion right ?

[wesm]

Reading to an Arrow table is zero-copy when the input is a memory map -- the use_threads is only for parallelizing conversions to pandas. Have you looked at parallelizing the equivalent conversions in R? https://issues.apache.org/jira/browse/ARROW-3315 (there is a companion issue for parallelizing the other direction)

[romainfrancois]

Not yet. at the moment focusing on the vignette and fixing things independent pr as needed

[wesm]

+1. Thanks @romainfrancois!

[wesm]

Build fails due to

0.78schecking Rd cross-references ... WARNING
Missing link or links in documentation object 'read_feather.Rd':
‘feather_table_reader’
See section 'Cross-references' in the 'Writing R Extensions' manual.
checking installed package size ... NOTE
installed size is 11.9Mb
sub-directories of 1Mb or more:
libs 10.0Mb
R 1.7Mbchecking S3 generic/method consistency ... NOTE
Found the following apparent S3 methods exported but not registered:
print.integer64 str.integer64
See section ‘Registering S3 methods’ in the ‘Writing R Extensions’
manual.

merge after fixing this

[wesm]

Thank you. merging