vspinu/unnest

Very fast spreading and stacking of nested data structures into R data.frames

Description

unnest is a zero-dependency R package for a very fast single-copy and single-pass unnesting of hierarchical data structures.

unnest(..) function takes a nested list as an input (JSON, XML etc.) and produces a data.frame according to an unnesting spec.

Installation

install.packages("unnest")
# or
devtools::install_github("vspinu/unnest")

unnest's Issues

`id` is not generated with atomic stacking

unnest(list(a = 1:3, b = 10:14), s(stack = "ix", as = "point", s(stack = "id")))
  ix point
1  a     1
2  a     2
3  a     3
4  b    10
5  b    11
6  b    12
7  b    13
8  b    14
> unnest(list(a = as.list(1:3), b = as.list(10:14)), s(stack = "ix", as = "point", s(stack = "id")))
  ix point point.id
1  a     1        1
2  a     2        2
3  a     3        3
4  b    10        1
5  b    11        2
6  b    12        3
7  b    13        4
8  b    14        5

Parallelize

Release v0.0.2 to CRAN.

Prepare for the release:

• Fix all milestone issues
• devtools::check()
• devtools::check_win_devel()
• devtools::check_win_release()
• rhub::check_for_cran()
• rhub::check_for_cran(check_args = c('--no-stop-on-test-error'), platforms = 'solaris-x86-patched')
• revdepcheck::revdep_check(num_workers = 8)
• Polish NEWS
• Polish pkgdown reference index
• Update dependencies
• Check revdeps
• Email reverse dependencies not to use it

Submit to CRAN:

• Change version in DESCRIPTION and NEWS.md (no-commit!)
• Update cran-comments.md
• devtools::submit_cran()
• Approve email

Wait for CRAN...

• Accepted!!!!
• usethis::use_github_release()
• usethis::use_dev_version()

"paste" processing is not handling list names correctly

> str(unnest(list(a = list(b = 1, c = 2)), s("a", process = "paste")))
'data.frame':	1 obs. of  1 variable:
 $ a: chr "1,2"

Implement joint (non-repetitive) unnesting

Stacking of empty vectors produces NAs

Browse[2]> unnest(list(a = 1:3, b = numeric()), s(stack = "ix", as = "point", s(stack = "id")))
  ix point
1  a     1
2  a     2
3  a     3
4  b    NA
Browse[2]> unnest(list(a = as.list(1:3), b = list()), s(stack = "ix", as = "point", s(stack = "id")))
  ix point point.id
1  a     1        1
2  a     2        2
3  a     3        3
4  b    NA       NA
Browse[2]> 

Add ability to reference paths starting from the root

Add `key` spec attribute to list keys first in the output data.frame