Manipulate shapefile attribute table using R

Question

I've posted this question on the GIS stack exchange, but it's not seeing much traffic.

I'm a GIS user who's been using R for stats for a few years, and I'm excited to see a lot of new GIS capabilities being released (raster, shapefiles, rgdal, etc).

I've done a lot of database and table manipulation in R, so being able to add and remove attributes from shapefiles is a powerful potential.

I'm hoping I'm just missing what's out there, but I can't seem to find a good method for adding or removing attributes from a shapefile attribute table.

Can anyone from Overflow respond to my GIS post? Or can I be advised how to get my list attribute table into a dataframe and back to replace the current attribute table?

Edit progress

Some progress since my original post:

This is getting at figuring out how to take my attribute table appart (.dbf), add stuff, and now I'm trying to put it back together to replace the original dbf.

>libary(raster); library(rgdal); library(shapefiles)
>shp<-shapefile(Shape)  # D.C. area airport polygons
>summary(shp)           #Shapefile properties

Object of class SpatialPointsDataFrame
Coordinates:
                min       max
coords.x1  281314.2  337904.7
coords.x2 4288867.0 4313507.0
Is projected: TRUE 
proj4string :
[+proj=utm +zone=18 +datum=WGS84 +units=m +no_defs +ellps=WGS84 +towgs84=0,0,0]
Number of points: 4
Data attributes:
    ObjectID        LOCID               NAME              FIELD              STATE            STATE_FIPS           ACAIS              TOT_ENP            TYPE          
 Min.   :134.0   Length:4           Length:4           Length:4           Length:4           Length:4           Length:4           Min.   :    271   Length:4          
 1st Qu.:242.8   Class :character   Class :character   Class :character   Class :character   Class :character   Class :character   1st Qu.:   3876   Class :character  
 Median :339.0   Mode  :character   Mode  :character   Mode  :character   Mode  :character   Mode  :character   Mode  :character   Median :3409113   Mode  :character  
 Mean   :310.0                                                                                                                     Mean   :3717251                     
 3rd Qu.:406.2                                                                                                                     3rd Qu.:7122488                     
 Max.   :428.0

 >shp.AT<-read.dbf(gsub(".shp", ".dbf", Shape), header=TRUE)    #Read in the attribute table from the .dbf
 > shp.AT       # First object in the dbf looks like an attribute table!
$dbf
  ObjectID LOCID                            NAME               FIELD    STATE STATE_FIPS ACAIS TOT_ENP          TYPE
1      134   ADW                     Andrews AFB                <NA> Maryland         24     Y    5078      Military
2      279   DCA             Washington National                <NA> Virginia         51     Y 6813148          <NA>
3      399   HEF               Manassas Regional Harry P Davis Field Virginia         51     Y     271      Regional
4      428   IAD Washington Dulles International                <NA> Virginia         51     Y 8050506 International

$header
$header$file.version
[1] 3

$header$file.year
[1] 113

$header$file.month
[1] 4

$header$file.day
[1] 12

$header$num.records
[1] 4

$header$header.length
[1] 321

$header$record.length
[1] 148

$header$fields
        NAME TYPE LENGTH DECIMAL
1   ObjectID    N     10       0
2      LOCID    C      5       0
3       NAME    C     45       0
4      FIELD    C     30       0
5      STATE    C     24       0
6 STATE_FIPS    C      2       0
7      ACAIS    C      1       0
8    TOT_ENP    N     11       0
9       TYPE    C     20       0

>shp.tab<-as.data.frame(shp.AT[1]) # Grab the first object of the .dbf as a data.frame

> shp.tab       # First list object
  dbf.ObjectID dbf.LOCID                        dbf.NAME           dbf.FIELD dbf.STATE dbf.STATE_FIPS dbf.ACAIS dbf.TOT_ENP      dbf.TYPE
1          134       ADW                     Andrews AFB                <NA>  Maryland             24         Y        5078      Military
2          279       DCA             Washington National                <NA>  Virginia             51         Y     6813148          <NA>
3          399       HEF               Manassas Regional Harry P Davis Field  Virginia             51         Y         271      Regional
4          428       IAD Washington Dulles International                <NA>  Virginia             51         Y     8050506 International 

> shp.tab$NewAT<-1:nrow(shp.tab) # Add my new attribute

> shp.tab # Added my new attribute, now to get this back into my shapefile
  dbf.ObjectID dbf.LOCID                        dbf.NAME           dbf.FIELD dbf.STATE dbf.STATE_FIPS dbf.ACAIS dbf.TOT_ENP      dbf.TYPE NewAT
1          134       ADW                     Andrews AFB                <NA>  Maryland             24         Y        5078      Military     1
2          279       DCA             Washington National                <NA>  Virginia             51         Y     6813148          <NA>     2
3          399       HEF               Manassas Regional Harry P Davis Field  Virginia             51         Y         271      Regional     3
4          428       IAD Washington Dulles International                <NA>  Virginia             51         Y     8050506 International     4

>write.dbf(shp.tab, gsub(".shp", ".dbf", Shape)) # Knew this wouldn't work, but demonstrate attempt to write this object as .dbf.
ERROR:
invalid subscript type 'list'

> shp.AT[1]<-shp.tab # Try replacing the old Object[1] with my new table containing the new attribute.

> shp.AT # The table portion fo the shp.AT is gone. No attributes.
$dbf
[1] 134 279 399 428

$header
$header$file.version
[1] 3

$header$file.year
[1] 113

$header$file.month
[1] 4

$header$file.day
[1] 12

$header$num.records
[1] 4

>write.dbf(shp.AT, gsub(".shp", ".dbf", Shape)) # If I go ahead and overwrite anyway...

My attributes are gone and replaced with an attribute table that contains a single field "dataframe". I recurse the script to again read the attribute table.

> shp.tab
  dataframe
1       134
2       279
3       399
4       428

So, I think I'm close. Can anyone help me get this back into the shapefile attribute table? Or, is there a way better method?

Thanks

Answer 1

I'm not sure I totally understand what you're trying to do. It looks like you just want to add a new column to the attribute table? If this is right, then just treat it like any dataframe.

library(rgdal)
dsn <- system.file("vectors", package = "rgdal")
shp<-readOGR(dsn = dsn, layer = 'cities')
shp$NewAT<-1:nrow(shp)

This works perfectly with a shapefile I have on my system. I typically rely on rgdal to read in my shapefiles, using the readOGR() function. I'm fairly certain the shapefile() function you were calling also calls rgdal.

Edited to add reproducable dataset.

Answer 2

Iam using R for GIS suff for several years now, usually in combination with QGIS. For manipulating attribute tables I usually save my shapes as CSV with the geometry as WKT (you will find the options in the QGIS save-dialog). In the next step I read all my csv(shapes) in R and do my statistics, joining, etc. Finally I write them back to the HDD and load them back into QGIS (no import-dialog needed just drag & drop) and save them as shapefiles.

HTH, Jo