Uboote in Regierung und Fraktion: glitches in s_attibutes("party") and s_attributes("parliamentary_group")

[mxi-hug]

I just ran into a couple of false s_attributes, both with party and parliamentary_group.

require(polmineR)
require(dplyr, exclude = "count")
use("GermaParl")

Issue 1: Parts of the Schröder I government seem to be rather conservative

# size of party in government by lp. CDU in government in lp 14/15 looks wrong. 
size("GERMAPARL", s_attribute = c("party", "role", "lp")) %>%
  filter(role == "government") 

# c'est qui?
corpus("GERMAPARL") %>%
  subset(party == "CDU" & 
           lp %in% c("14","15") & 
           role == "government") %>%
  s_attributes(c("speaker", "date", "role", "party")) %>% 
  group_by(speaker, party) %>%
  summarise(n = n())

Issue 2: parliamentary_group occasionally differs from party

# Verhältnis von Partei zu Fraktion:
size("GERMAPARL", s_attribute = c("party", "parliamentary_group"))

# Example
# Wer ist das Uboot?
corpus("GERMAPARL") %>%
  subset(party == "FDP" & 
           parliamentary_group == "GRUENE") %>%
  s_attributes(c("speaker", "date", "role", "party", "parliamentary_group"))

# Gegenprobe. Ist Solms gar ein Grüner? Nur manchmal.
corpus("GERMAPARL") %>%
  subset(speaker == "Hermann Otto Solms") %>%
  size(c("speaker", "parliamentary_group"))

This is obviously no complete list of all speakers but the issue should be clear.

I'm currently using GERMAPARL version 1.0.6