[R-sig-Geo] Create SpatialLinesDataFrame

Wed Feb 17 08:25:56 CET 2010

On Wed, 17 Feb 2010, Michael Sumner wrote:

>>> flop<-Lines(Line(list(rbind(as.numeric(temp[1,7:8]),
>> as.numeric(temp[1,9:10])))), ID=row.names(temp))

I think that there is a misplacement too:

flop <- Lines(list(Line(rbind(as.numeric(temp[1,7:8]),
   as.numeric(temp[1,9:10])))), ID=row.names(temp))

looks better - Lines() takes a list of Line objects as its first argument.

Roger

>>> SpatialLines(flop)
>> Error in as.list.default(X) :
>>  no method for coercing this S4 class to a vector
>>
>
> You need to put "flop" in a list of its own, as SpatialLines is a
> collection class that may contain more than one "Lines" (think of
> multi-branched line sequences that may not be connected, but exist as
> a single entity/feature - a "Line" is a matrix of a coordinates that
> are connected in sequence, a "Lines" is a collection of those, and a
> SpatialLines is a collection of Lines - each Lines will have a row in
> the attribute table). So SpatialLines(list(flop)) should work.
>
> See in my code how I start with a generic empty list, then populate it
> with many Lines:
>
> ranlines <- list()
>
> for (irow in 1:nrow(temp)) {
>       ranlines[[irow]] <- Lines(Line
>
>
>> Why is this?  Also to export my SLDF as a shapefile I would normally use the
>> following
>>> writeSpatialShape(ranslines, "C:/GIS/ranslines")
>>
>> ...but this is different from what Michael has proposed.
>>> writeOGR(sldf, ".", "ranlines", "ESRI Shapefile")
>>
>> Is one method better than the other?
>>
>
> I use the rgdal package as it was my first introduction to read/write
> for GIS data in R, and it is very general. It can be hard to install
> as there are 3rd-party dependencies even for the base package (very
> easy on Windows with a binary/zip package), though when you compile
> from source you have greater control and the potental for adding new
> drivers (formats) over and above the base ones. There may be important
> differences between the way shapefiles are written but it's personal
> preference and ease of use/installation as far as I can see.
>
> cheers, Mike.
>
> On Wed, Feb 17, 2010 at 10:28 AM, Tyler Dean Rudolph
> <tylerdeanrudolph at gmail.com> wrote:
>> Much thanks to all for these useful and informative responses.
>>
>> I had come across the book in my searches but I had not stumbled onto its
>> homepage with the additional resources Roger has provided, which are
>> indispensable.  Perhaps to circumvent the potential conflict of interest you
>> could add a URL to your personal signature referring interested parties to a
>> website containing information on your related books and projects?  This to
>> me seems entirely reasonable as it does not solicit interest, it merely
>> provides access to that information.
>>
>> I realize that have a different way of accessing the package vignettes (or
>> so I thought: http://cran.r-project.org/web/packages/sp/sp.pdf) so I haven't
>> necessarily been working with all the available information either.
>>
>> I will keep Robert's handy conversion code nearby, which promises to make
>> creating sp objects quick and easy.
>>
>> Thank you to Michael for his solution and encouragement.  I gathered that
>> the "ID" argument was critical to successfully passing the SpatialLines
>> command.  I considered doing a "for" loop, but when the following failed to
>> work with one mere record I decided there was no solution there to be found.
>>
>>> flop<-Lines(Line(list(rbind(as.numeric(temp[1,7:8]),
>> as.numeric(temp[1,9:10])))), ID=row.names(temp))
>>> SpatialLines(flop)
>> Error in as.list.default(X) :
>>  no method for coercing this S4 class to a vector
>>
>> Why is this?  Also to export my SLDF as a shapefile I would normally use the
>> following
>>> writeSpatialShape(ranslines, "C:/GIS/ranslines")
>>
>> ...but this is different from what Michael has proposed.
>>> writeOGR(sldf, ".", "ranlines", "ESRI Shapefile")
>>
>> Is one method better than the other?
>>
>> Tyler
>>
>>
>> On Tue, Feb 16, 2010 at 4:06 AM, Roger Bivand <Roger.Bivand at nhh.no> wrote:
>>
>>> On Mon, 15 Feb 2010, Tyler Dean Rudolph wrote:
>>>
>>>  I have searched extensively for some good examples or documentation on how
>>>> to create a SpatialLinesDataFrame and for the life of me have found zilch.
>>>> This does not mean there is no such thing available, it just means that
>>>> wherever it is it is not readily accessible through the circuitous
>>>> pathways
>>>> of R help documentation or via a simple web search.
>>>>
>>>
>>> I have discussed with Achim Zeilis, who faces the same problem, the ethical
>>> appropriateness of mentioning our book on the Spatial Task View, and we've
>>> found until now that we shouldn't mention our own books. If opinion is that
>>> we should, and because they are in a number of university libraries, I can
>>> do that, and encourage Achim to do the same for Econometrics.
>>>
>>> If you visit the book website (http://www.asdar-book.org), you will find
>>> links to code showing how to do this (chapter 2). In addition, under
>>> "Additional materials", you find links to course materials, especially to a
>>> course given at Imperial College in 2007, in which a good deal of these
>>> questions are covered (see "Representing Spatial Data").
>>>
>>> Because the classes were designed as interface containers, most users do
>>> not need to know how to construct them from scratch, typically reading in
>>> data or coercing from other classes of objects. While the help pages could
>>> certainly be improved, we do provide a vignette that ships with the package,
>>> and where section 6 covers Line objects etc. See:
>>>
>>> http://cran.r-project.org/web/packages/sp/vignettes/sp.pdf
>>>
>>> or:
>>>
>>> vignette("sp", package="sp")
>>>
>>> in R.
>>>
>>> I'm sure that Mike's answer resolves your immediate problem - this is more
>>> of an answer to your general question about where to look for information
>>> about sp classes and how to make and use them.
>>>
>>> Hope this helps,
>>>
>>> Roger
>>>
>>>  As someone who
>>>> struggles daily to achieve highly technical things with scarcely the savvy
>>>> required, I wish there were more (or any?) examples in the documentation
>>>> for
>>>> most of the spatial packages available.  (I also wish that managing and
>>>> manipulating Spatial objects was about 65% less complicated, but for the
>>>> moment let's start with this)...
>>>>
>>>> I have a data.frame (temp) containing endpoints of animal trajectories
>>>> with
>>>> which I would like to create a SpatialLinesDataFrame for export into a
>>>> GIS.
>>>>
>>>>  temp
>>>>>
>>>>      id                 day       x        y        X2       Y2
>>>> 1 2002007 2005-04-04 12:01:01 -300358 748992.6 -300450.8 748764.8
>>>> 2 2002007 2005-04-04 12:01:01 -300358 748992.6 -301694.2 751728.4
>>>> 3 2002007 2005-04-04 12:01:01 -300358 748992.6 -299983.1 749528.8
>>>> 4 2002007 2005-04-04 12:01:01 -300358 748992.6 -300569.4 749640.7
>>>> 5 2002007 2005-04-04 12:01:01 -300358 748992.6 -300439.0 748308.8
>>>>
>>>> I seem reasonably able to create a Lines object (though I can't get the ID
>>>> argument to work for me where there are multiple records involved)....
>>>>
>>>>  ranlines<-apply(temp, 1,  function(x)
>>>>>
>>>> Lines(Line(list(rbind(as.numeric(x[7:8]), as.numeric(x[9:10]))))))
>>>>
>>>>  str(ranlines[[1]])
>>>>>
>>>> Formal class 'Lines' [package "sp"] with 2 slots
>>>>  ..@ Lines:List of 1
>>>>  .. ..$ :Formal class 'Line' [package "sp"] with 1 slots
>>>>  .. .. .. ..@ coords: num [1:2, 1:2] -300358 -300451 748993 748765
>>>>  .. .. .. .. ..- attr(*, "dimnames")=List of 2
>>>>  .. .. .. .. .. ..$ : NULL
>>>>  .. .. .. .. .. ..$ : chr [1:2] "X1" "X2"
>>>>  ..@ ID   : chr NA
>>>>
>>>> Am I at least on the right track?   Unfortunately now I'm clueless as to
>>>> what to do next because apparently Lines cannot be coerced to
>>>> SpatialLines,
>>>> and while I am normally able to figure my way out of things there is just
>>>> nothing intuitive about this as far as I can tell.  I have tried to create
>>>> SLDFs numerous times in the past and I have always been discouraged from
>>>> continuing.  I can create a psp object in Spatstat but I am equally unable
>>>> to convert that into a SLDF shape.  Terribly sorry about the rant but it's
>>>> difficulties like these that prevent a lot of people more clueless than me
>>>> from ever getting anywhere with R and more specifically the spatial
>>>> stuff....
>>>>
>>>> Tyler
>>>>
>>>>        [[alternative HTML version deleted]]
>>>>
>>>> _______________________________________________
>>>> R-sig-Geo mailing list
>>>> R-sig-Geo at stat.math.ethz.ch
>>>> https://stat.ethz.ch/mailman/listinfo/r-sig-geo
>>>>
>>>>
>>> --
>>> Roger Bivand
>>> Economic Geography Section, Department of Economics, Norwegian School of
>>> Economics and Business Administration, Helleveien 30, N-5045 Bergen,
>>> Norway. voice: +47 55 95 93 55; fax +47 55 95 95 43
>>> e-mail: Roger.Bivand at nhh.no
>>>
>>>
>>
>>        [[alternative HTML version deleted]]
>>
>> _______________________________________________
>> R-sig-Geo mailing list
>> R-sig-Geo at stat.math.ethz.ch
>> https://stat.ethz.ch/mailman/listinfo/r-sig-geo
>>
>
> _______________________________________________
> R-sig-Geo mailing list
> R-sig-Geo at stat.math.ethz.ch
> https://stat.ethz.ch/mailman/listinfo/r-sig-geo

-- 
Roger Bivand
Economic Geography Section, Department of Economics, Norwegian School of
Economics and Business Administration, Helleveien 30, N-5045 Bergen,
Norway. voice: +47 55 95 93 55; fax +47 55 95 95 43
e-mail: Roger.Bivand at nhh.no