New sync()

[maurolepore]

This commit adds pick() which is the basis for sync() -- a general function to apply a function to one key dataframe in a list and propagate the indices to all other elements.

[coveralls]

Pull Request Test Coverage Report for Build 596

• 0 of 0 changed or added relevant lines in 0 files are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage increased (+0.8%) to 87.165%

---

Totals
Change from base Build 595:	0.8%
Covered Lines:	618
Relevant Lines:	709

---

💛 - Coveralls

[maurolepore]

Ideas:

data /
  sync(key = NULL) /
  filter() /
  arrange(1) /
  unsync()

Code within sync() has the sync class with special behaviour. Whatever happens on rows is applyed to the key -- if not null -- then propagated to the other elements of the list. Maybe it understands only the main 5-dplyr-verbs.

Another way

data /
  sync(exression)

For example

data /
  sync(filter(this_column == "this_value"), key = "this_column")

or

data /
  sync(this_key_col, filter)(this_col == "this_value")

• Add in a package, {census} that creates objects of class census (or similar)
• trunk may refer to a something analog to a tree or git master branch
• branch may refer to any branch/stem (master or not)
• Class census ~ tree and stem
• Subclass normalized ~ tree
• pick.census prints as nested datafame.
• How do other networks store and analyze their census data?
• {tsibble} may have something useful. In particular slide() and tyle() look good.

[maurolepore]

Relates, is the {rtbl} project.

Also consider a map_on() and map_off():

data /
  nest() /
  map_on(data) /
  filter(this_col == "this_value") /
  map_off()
  unnest()