Remove empty polygons from shapefile in R

Question

I have a shapefile witch contains several empty polygons. I´m extracting values from raster file based in some polygons of this shapefile. But my loop stops at every empty polygon. Simple solution would be to remove NAs with na.omit(shape@data) but this deletes some xy values as well.

Here is my code. The loop stops once it finds a empty polygon.

all.mun <- NULL
  for (j in mun.codes)
  {
    mun       <- mun.br[mun.br$CODIGO_MUN == j, ] # isolate municipality
    suit.mun  <- extract(suit, mun)               # extract suits from the municipality
    suit.mun  <- do.call("c", suit.mun)           # convert to data frame no matter how many lists the function extract::raster returns
    suit.mean <- mean(suit.mun)                   # mean suitabilitie of the site
    all.mun   <- rbind(all.mun, suit.mean)        # saving result by observation, i.e municipalitie
  }

mun.condes is a vector with a list of 6 digit municipalities codes. mun.br is the shapefile of the municipalities

> names(mun.br)
 [1] "ID"         "NOMEUF"     "REGIAO"     "CODMESO"    "NOMEMESO"   "CODMICRO"  
 [7] "NOMEMICRO"  "NOME_1"     "NOME_2"     "CODIGO_UF"  "CODIGO_MUN" "AREA_1997" 

Here is the error message.

Error in y[i, ] : 
  cannot get a slot ("Polygons") from an object of type "NULL"

Answer 1

As mentioned by Spacedman, it is much easier to address your issue with reproducible data. To make sure you address both the possible concern of emptiness and not matching CODIGO_MUN, I would try

library(raster) # make sure next time that you specify which libraries you are using, I am assuming these suffice
library(rgdal)
all.mun <- NULL
  for (j in levels(as.factor(mun.br$CODIGO_MUN))) # make sure all 'j' match CODIGO_MUN's values
  {
     if (!identical(mun.br[mun.br$CODIGO_MUN == j, ]@polygons, NULL)) {  # if polygon is not NULL
     # if (length(mun.br[mun.br$CODIGO_MUN == j, ]@polygons)>0) { # alternative if conditional if the former does not work
        mun       <- mun.br[mun.br$CODIGO_MUN == j, ] # isolate municipality
        suit.mun  <- extract(suit, mun)               # extract suits from the municipality
        suit.mun  <- do.call("c", suit.mun)           # convert to data frame no matter how many lists the function extract::raster returns
        suit.mean <- mean(suit.mun)                   # mean suitabilitie of the site
        all.mun   <- rbind(all.mun, suit.mean)        # saving result by observation, i.e municipalitie
     }
  }

Note that I am assuming that an object without polygon has a polygon slot equivalent to NULL. You can try identical(...,NA) or identical(...,integer(0)) in case the former does not work, or the one I left commented.