Make dfm_match() work with padding

[koheiw]

In koheiw/newsmap#39, I noticed that dfm_match() crashes when features contain "".

toks <- tokens("a b !", padding = TRUE, remove_punct = TRUE)
dfmt <- dfm(toks)
dfm_match(dfm("a b c"), "") # error

The root cause is actually [.dfm

dfmt[,""] #  error
dfm_select(dfmt, "") # works

This is a deeper issue that requires careful redesigning of how we handle paddings, because R treats empty names as a special case according to ?names():

The name "" is special: it is used to indicate that there is no name associated with an element of a (atomic or generic) vector. Subscripting by "" will match nothing (not even elements which have no name).

[codecov]

Codecov Report

Merging #1960 into master will not change coverage.
The diff coverage is 100.00%.

@@           Coverage Diff           @@
##           master    #1960   +/-   ##
=======================================
  Coverage   91.26%   91.26%           
=======================================
  Files         101      101           
  Lines        6388     6388           
=======================================
  Hits         5830     5830           
  Misses        558      558