Output unification from geocode

[kadyb]

Depending on the given object, there are returned lists with different attributes as a result. It would look better if it was unified. Note, a single list or multiple lists (if there are n of the same objects in different locations) can be returned.

library("rgugik")

# place
lapply(geocode(address = "Marki")[1], names)
#> $`1`
#> [1] "city"         "teryt"        "simc"         "voivodeship" 
#> [5] "county"       "commune"      "x"            "y"           
#> [9] "geometry_wkt" "accuracy"     "id"           "jednostka" 

# place and street
names(geocode(address = "Marki, Andersa")) # IT'S LINE, NOT POINT
#> [1] "street"       "teryt"        "simc"         "ulic"        
#> [5] "city"         "x"            "y"            "geometry_wkt"
#> [9] "accuracy"     "id"           "jednostka"   

# place, street and house number
names(geocode(address = "Marki, Andersa 1"))
#> [1] "city"            "citypart"        "street"         
#> [4] "number"          "teryt"           "simc"           
#> [7] "ulic"            "code"            "jednostka"      
#> [10] "x"               "y"               "geometry_wkt"   
#> [13] "accuracy"        "city_accuracy"   "street_accuracy"
#> [16] "id"   

# place and house number
names(geocode(address = "Królewskie Brzeziny 13"))
#> [1] "city"          "citypart"      "street"       
#> [4] "number"        "teryt"         "simc"         
#> [7] "ulic"          "code"          "jednostka"    
#> [10] "x"             "y"             "geometry_wkt" 
#> [13] "accuracy"      "city_accuracy" "id"  

# physiographic object 1
names(geocode(geoname = "Las Mierzei"))
#> [1] "name"         "genitive"     "voivodeship"  "county"      
#> [5] "commune"      "class"        "type"         "status"      
#> [9] "geometry_wkt" "x"            "y"            "accuracy"  

# physiographic object 2
lapply(geocode(geoname = "Jeziorak")[1], names)
#> [1] "name"         "genitive"     "voivodeship"  "county"      
#> [5] "commune"      "class"        "type"         "status"      
#> [9] "notes"        "geometry_wkt" "x"            "y"           
#> [13] "accuracy"

In case "road" or "rail crossing", output is OK.

[Nowosad]

@kadyb I can think of two solutions:

1. Find one set of common variables (I would guess ~20 based on your examples above). Create an object template with these ~20 variables (columns) and next fill them with downloaded data. In case when some query does not return a variable - the variable stays NA.
2. Split possible results into several groups (as above). Treat each group independently.

What do you think about this issue?

[kadyb]

To information: We currently have a lot of duplicate code, so we should create some helper function.

[kadyb]

BTW. is it possible that geocodePL_get() may return sf object instead of list? That would be super useful for speeding up the processing and merging with other data?

Fixed in https://github.com/kadyb/rgugik/pull/43.

[BERENZ]

Concerning this issue, maybe we could leave this for the user? We do not know which columns they will use or find useful for their analysis.

[kadyb]

I think we can safely remove the meaningless columns: x, y, id, jednostka, accuracy, city_accuracy, street_accuracy