Missing column names in return value of s_attributes() for subcorpora with single cpos pair

[ChristophLeonhardt]

Using GERMAPARLMINI in polmineR v0.8.8.9010 I observed the following unexpected behaviour:

When applying s_attributes() with more than one s-attribute to subcorpora with only one cpos pair, the returned data.table does not have correct column names.

Take the following example:

library(polmineR)
use("polmineR")

corpus("GERMAPARLMINI") |>
  subset(speaker == "Norbert Lammert") |>
  subset(party == "CDU_CSU") |>
  s_attributes(c("speaker", "party"))

Expected would be a data.table with the column names "speaker" and "party". But the return value is a data.table with the column names "V1" and "V2".

After looking at s_attributes.R, I think the issue is that subcorpora with a single cpos pair (or partitions which behave the same) are treated differently from subcorpora with more than one cpos pair, thus losing the name of the s-attributes in the process. This is the part of the code I looked at:

polmineR/R/s_attributes.R

Lines 287 to 290 in db73937

	tab <- if (nrow(.Object@cpos) > 1L)
	data.table(tab_data)
	else
	data.table(matrix(tab_data, nrow = 1L))

If there is only one cpos pair, the tab_data object, which still contains the correct names, is turned into a matrix object first which does not have the column names anymore.

Using the dimnames argument of matrix() might be a solution here since tab_data could provide column names.