fixed duplicates in FAO-Commodities and in `name_to_rgn_id()`

[bbest]

(OHI-Science/ohicore#34)

Renamed cbind_rgn() to name_to_rgn_id() (in src/R/ohi_clean_fxns.R), since major overhaul of duplicate handling. Now the function drops the original fld_name and returns rgn_id with any duplicates collapsed by collapse_fxn. For instance with the FAO-Commodities, the following duplicates were collapsed:

                         rgn_id
  country                  73 140 186 209 244
    China, Hong Kong SAR    0   0   0 180   0  # China to be summed?
    China, Macao SAR        0   0   0 144   0  # China to be summed?
    Martinique              0  36   0   0   0  # Guadeloupe | Martinique => Guadeloupe and Martinique region
    Serbia and Montenegro   0   0  36   0   0  # Serbia | Montenegro => Serbia and Montenegro region
    Un. Sov. Soc. Rep.     36   0   0   0   0  # Un. Sov. Soc. Rep. > Russian Federation => Russia

Other fixes:

• Dropped from Netherlands Antilles expansion since already in data:
  • Aruba
  • Curacao
• Added to src/LookupTables/rgn_eez_v2013a_synonyms.csv:
  • Cabo Verde (ohi_region 41)
  • Ethiopia PDR (landlocked)

[jules32]

Hi @bbest,

I'm editing a few lines in name_to_rgn_id.r slightly because using first returns only the very first entry, so that the output variable rgn has only one entry: rgn_id = 1 for Cocos Islands. I'm not sure what first was intending to do? I have made sure this isn't a matter of plyr instead of dplyr, but I will remove first.

[jules32]

Sorry I linked those lines to the current version instead of the historical record. This link shows what those lines had been, when first was included.

[jules32]

Hi again @bbest,

It also seems that the group_by function is not working as we want it to. This bit will only return two columns:

          rgn_name rgn_type
1    Cocos Islands      eez
2 Christmas Island      eez
3   Norfolk Island      eez
4 Macquarie Island      eez
5    New Caledonia      eez
6          Vanuatu      eez

so I am changing it to include rgn_id in the summarize() part:

 group_by(rgn_id) %.% 
    summarize(
      rgn_id,
      rgn_name = rgn_name,
      rgn_type = rgn_type) %.%
    ungroup()

I will keep you posted.

[bbest]

I like the problem labeling: "the Vanuatu issue" like "the Eritrea issue". There's something amiss here, since anytime you use group_by(rgn_id), it should always return rgn_id field after running any summarize() operation. The fact that it doesn't and you have to explicitly include rgn_id makes me suspicious as to whether it's properly grouping in the first place. I again suspect something afoul with package sequence loading, ie library(dplyr) should always come after library(reshape) and library(plyr).