w to
zonal<SpatRaster,SpatRaster> to compute weighted
meanszonal<SpatRaster,SpatVector> methodclamp_ts methodsplit<SpatVector>. #979 by
srfallReleased 2023-01-17
shade is now vectorized for arguments
angle and direction to facilitate generating
multiple hillshades that can be combined for a better result #948 by Jürgen
NiedballasharedPaths now uses spatial indices #960 by Jeff
Hansonpredict has better support for models such as ranger
that do not return anything for missing values #968 by Alex
IlichwriteCDF now supports writing yearly time steps #926 by Andrea
ManicaReleased 2022-12-02
roll<SpatRaster> method for rolling (moving)
average and other rolling functionsnoNA<SpatRaster> method to identify cells that
are not NA (across layers)rangeFill<SpatRaster> methodexhaustive to
spatSample<SpatRaster> for large sparse rasters. #905 by
PetiteTong.focalPairs and focalReg can now use the
values in custom windows as weights. #907 by Fabian
Fischer.focalReg now has additional arugment “intercept=TRUE”.
#916 by
Jordan Adamsoncrs(x, warn=TRUE)<- now emits a warning about the
difference between transforming and setting a crs when x already had a
crs. #897 by
Márcia Barbosa.writeRaster #900 by Kyle
Davidcrosstab now shows the labels names for a categorical
SpatRaster. 895 by Derek
Corcoran BarriosmakeTiles can now take a SpatVector to define the
tiles. 920 by
Tristan GoodbodyfocalPairs and focalReg now work for
custom windows #907 by Fabian
Fischerplot<SpatVector> was not
working properly. #906 by Márcia
Barbosa.time<- with time-step “years” could not handle
negative years. #911 by Andrea
Manicawrap/unwrap (and by extension
saveRDS/readRDS) did not handle categorical
rasters well #912 by
Christine Anderson.interpIDW failed with GDAL 3.6 #910 by Roger
BivandspatSample with strata bug fix “unable to find an
inherited method for function ‘trim’” #919 by Alfredo
AscanioReleased 2022-11-18
[ and [<- for SpatRaster now have a
third index k for subsetting or assigning values by
layeranyNA and allNA for SpatRasterunwrap to restore a PackedSpatVector or
PackedSpatRasterrasterizeWin method for rasterization with a moving
window (circle, ellipse, rectangle, buffer)interpIDW method for inverse-distance-weighted
interpolation of points with a moving windowinterpNear method for nearest neighbor interpolation of
points with a moving windowviewshed method for SpatRasterupdate method for SpatRaster to write new names or a
new extent or crs to an existing raster file.sieve filter for SpatRastersegments=FALSE to
disagg<SpatRaster>sprc<character> method to create a
SpatRasterCollection from a file with subdatasetsgraticule function to create a SpatGraticule and
related methods plot<SpatGraticule> and
crop<SpatGraticule>elongate method for SpatVector linesmosaic and merge<SpatRaster>
#577 by
Jean-Romainwrap<SpatRaster> now uses file references if the
data is deemed to large to all load into memory. #801 by
Jean-RomainreadRDS and unserialize now return a
SpatRaster or SpatVector (instead of a PackedSpat*)clamp can now take low and high values for each
layerpax argument in plot now provides more
control over what to draw on each axis via paramters side,
tick and labpax argument in plot now has argument
retro to use a sexagesimal notation of degreesextend has a new argument fill=NAcombining SpatRasters with
different CRSs. #818 by Andrew
MarxplotRGB now accounts for the value of zlim when
stretching; allowing to use the same coloring scheme across SpatRasters
#810 by
Agustin Lobo.spin is now vectorizedError in x$.self$finalize() : attempt to apply non-function
is now suppressed in most cases. #218 by Charlie
Joey Hadley. This problem should go away altogether when a new version
of “Rcpp” is released (ETA Jan 2023) thanks to a fix by Kevin Ushey #30spatSample with na.rm and SpatRasters with
multiple layers did not work. #800 by Andrea
Manicaadjacent<SpatRaster> with
pairs=TRUE, include=TRUE ignored include=TRUE
#808 by
Joseph Lewisrasterize did not accept “NA” as value for updating #809 by Márcia
Barbosaextract with a perfectly vertical or horizontal line
failed in some cases #823 by Dimitri
Falkwrap<SpatVector> failed if there was a single
point geometry #815 by Patrick
Schaeferextract<SpatRaster> with
weights=TRUE did not return values #814 by Jean-Luc
Dupouey.x[["newname"]] <- r for SpatRasters x
and r did not work #795 by Jim
ShadyReleased 2022-09-10
droplevels for SpatRaster. [#757] by Rodolfo
Jaffe.normalize.longitude for SpatVector.scoff to get and scoff<- to set the
scale (gain) and offset of a SpatRaster.raw=FALSE to
extract<SpatRaster> [#776] by Thomas Roh.as.data.frame now takes na.rm=NA to only
remove rows that are NA for all layers. The default value changed from
TRUE to NA. [#792] by Ed Carnelldistance<SpatRaster> has new arguments “target”
and “exclude”. [#560] by Bernardo Brandão Niebuhrsparse=FALSE for
`relate<SpatVector,SpatVector>.usenames=FALSE for
lapp<SpatRasterDataset> [#793] by Colin Brust.vect<character> now reports that a file is
non-existent [#784] by John Baumsrelate [#716] by Krzysztof Dybafocal3D now checks if all the window’s dimensions are
odd [#772] by Neander Marcel Hemingvrt is no longer constrained by the maximum number of
files that can be opened [#780] by 8Ginette8weighted.mean crashed with numeric weights and
na.rm=TRUE [#777] by David Holstiusproject<SpatRaster> did not consider an extent
that was set by the user [#775] by Philippe MassicottefocalCor failed for large rasters [#607] by John
Clarkfocal with expand=TRUE was prone to run
out of memory [#610] by Nathan Elliottcrop<SpatVector> did not work well when the
second argument were points or lines [#782] by Márcia Barbosaadjacent with pairs=TRUE now respects the
include=TRUE argument [808] by Joseph LewisReleased 2022-08-07
blocks to guide reading raster data in chunks.
[#748] by John Baumsextract were simplified. [#736] by
François Rousseufocal where not correct if the input
SpatRaster had multiple layers and a “custom” function. [#727] by
Jean-Luc Dupouey.plot<SpatRaster> did not honor argument
legend=FALSE. [#738] by Grzegorz Sapijaszkoexpanse failed when processing in chunks [#741] by
Gareth Daviescrop<SpatRaster,SpatExtent> with argument
snap="out" could lead to a crash if the extent was beyond
the SpatRaster. [#740] by Mauricio Zambrano-BigiariniReleased 2022-07-25
subst no longer uses values that it changed earlier on.
[#639] by Paul Smithas.points<SpatRaster> could return wrong factor
labels. [#640] by Attilio Beninimask<SpatRaster,SpatVector> crashed when the
results were written to disk. [#646] by Monika Anna Tomaszewskaextract<SpatRaster,SpatVector(points)>(xy=TRUE)
returned the locations of the points, not the xy-coordinates of the
cells. [#650] by Ward Fonteynwrap<SpatRaster> did not return the correct
labels for some categorical rasters. [#652] by Jakub Nowosadsubst can now substitute the values from multiple input
layers with a single output value (layer)subset<SpatVector> now behaves like
subset<data.frame> [#648] by Andrew Gene Brownfreq now provides the labels of factors, even if
bylayer=FALSE. It now always returns a
data.frame (it used to return a matrix in some
cases. [#687] by Rodolfo Jaffédisagg and aggregate now return a warning
instead of an error when using a (dis)aggregation factor of 1.[#684] by
Justin Fain.project crashed when erroneously projecting raster data
from one celestial body to another [#688] by Mike Sumnerplot. [673] by Jakub Nowosad.extract(method="bilinear"). [693] by swooping-magpieas.raster to
unique<SpatRaster> to create a categorical raster
with the unique combinations in the layers of the input raster. The
default for argument na.rm was changed to
FALSEsort<SpatRaster> to sort cell values across
layers.has.colors and has.RGB for SpatRastercover can now combine categorical rastersconcats to combine the levels of two SpatRaster into
new categories [663] by Alex Ilichzonal<SpatVector,SpatVector> method to aggregate
SpatVector attributes by polygonsReleased 2022-06-09
extract. [#580] by André M. Bellvéweighted.mean did not adjust the weights if there were
NAs in the values. [#574] by Lars Dalbystretch option in plotRGB [#550] by
Agustin LobospatSample with cells=TRUE failed for
planar data [#544] by Benjamin MisiukcompareGeom(x, y, stopOnError=FALSE) did not remove the
error messages stored in x leading to unexpected warnings
later on. [#568] by David Hofmann.layerCor does not create temp files anymore. [#551] by
Christine Andersonres and origin to
project<SpatRaster> method. [#596] by Alex Ilichinside=TRUE to centroids to
get a centroid-like point that is guaranteed to be on the geometry
(“point on surface”). [#588] by Márcia Barbosakeepgeom=FALSE to
vect<data.frame> that allows setting (keeping) the
geometry as an attribute. [#586] by Márcia BarbosasaveRDS and serialize methods for
SpatRaster and SpatVector. [#549] by Andrei MîrțxFromCol and yFromCol now have a
<SpatRaster,missing> method. [#583] by Michael
Sumner.svc<sf> method to deal with GeometryCollection
types. [#585] by Sarah Endicottas.points<SpatRaster> and
as.polygons<SpatRaster> have a new argument
na.all=FALSE that affects the interpretation of
na.rm. [#548] by Jean-Luc Dupouey.setGDALconfig and getGDALconfig to set
GDAL configuration options. [#608] by Erik Bolch.circular to rapp to allow the
start to be after the end (for if layers represent days of the
year)costDistance<SpatRaster>where.min and where.max for
SpatRaster to get the cell numbers for the extreme values
in a SpatRaster.emptyGeoms<SpatVector> to get the
indices of empty (null) geometriesrasterizeGeom to rasterize the geometry
count or the area of (small) polygons or the length of lines.not.na for SpatRaster which is
a shortcut for !is.na(x).as.list implemented for
<SpatRasterDataset>.sources implemented for
<SpatRasterDataset>, <SpatVector>
and <SpatVectorProxy> [#638] by Andrew Gene
BrownReleased 2022-02-17
writeVector and vect now work with GPGK if
the path has non-ascii characters [#518]predict with cores > 1
and more than one output variable were garbledzonal dropped category names when using an external (R)
function [#527] by Jakub Nowosadxy=TRUE in as.data.frame normalized
the names [#538] by Kodi Arferoptions to vrt [#629] by
Monika Tomaszewska.makeTiles has new arguments extend and
na.rm [#520] by by L. Dalbyproject<SpatRaster> now uses nearest neighbor as
default method for RGB rastersna.rm=TRUE to unique. [#561]
by Matthieu StiglerReleased 2022-01-30
app<SpatRasterDataset> ignored the filename.
[#498] by jszhaovect<data.frame> failed silently if xy
coordinates were integers [#496] by Márcia Barbosaaggregate<SpatRaster> was malformed
when nrow(x) %% fact != 0. [#492] by Jean-François
BourdonNAs in SpatVector attributes where only
recognized on Windows [#491] by Márcia Barbosaplot<SpatVector> failed when using a character
variable with many unique values. [#489] by Márcia Barbosarotate failed on large files. Reported by Ujjawal
Singhcrds replicated the coordinates [#504] by Murray
Effordas.data.frame<SpatRaster> returned integers if
the file stored values as integers, even if there was a scale/offset
that creates decimal numbers [#509] by Kodi Arferproject opened the input raster file in read/write mode
intead of read mode. That did not work with files that cannot be
updated.distance, gridDistance,
direction and patches now process all layers
of the input SpatRaster. [#503] by Chris Haak()<- methods.
in-place updating available with set. methods such as
set.names and set.values. [#493] by Jean
Romain and [#511] by Bryan Fuentesaggregate<SpatRaster> now accepts functions that
return more than one value per aggregated cellwriteVector has new argument insert to add
a layer to an existing file (e.g. GPKG).method="weights" for
spatSample<SpatRaster>mask<SpatVector,SpatVector> method to select
intersecting geometriesis.relatedvalues<SpatRaster> has new option
na.rm=TRUE. [#490] by Henk HarmsenSpatVectorProxy to provide access to large
vector databases that cannot or should not be read into memory in its
entirety.proxy=FALSE to vect to create
a SpatVectorProxy objectquery<SpatVectorProxy> to extract
parts of a SpatVectorProxyvector_layers that returns, and can delete,
vector format layers from a database/file such as GPKGTo avoid name clashes with tidyverse
For consistency
Released 2022-01-13
setValues and init failed (or even crashed
R) when using a single value on a largish raster. [#414]sfc to SpatVector lost the
crs. [#415] by Jean-Luc Dupoueybuffer on a SpatRaster with no values caused a crash
[#416] by Sebastian BrinkmannwriteVector now assumes “traditional GIS order”
(long/lat) if the CRS specifies lat/long. [#333]( by Agustin Lobomain was ignored in density when
using a single layer SpatRaster [#424] by dvictorimin and
mean, when used with multiple SpatRasters and numbers,
ignored additional SpatRasters [#426] by Zhuonan Wangclassify with right=FALSE ignored
include.lowest=TRUE [#442] by Alex Ilichpatches now combines patches that connect across the
data line [#366] by Hirschtpatches(directions=8) now connects in NE/SW direction
[#451] by Jean-François Bourdon.centroids now considers cases where SpatVector parts
are nearest to each other when crossing the date line in stead of the
zero-meridian [#366] by Hirschtterrain created empty (NA) rows between
chunks used for processing large rasters. [#453] by Robert Ritson.inset did not draw the “box” correctly. [#457] by
Márcia Barbosaas.lines now works with a points SpatVector [#464] by
Márcia Barbosavalues(x)<- now accepts (hex coded) colors as
valuesfocal now wraps around the dateline like raster::focal
[#242] by Alexander Marbleraggregate now does not show a progress bar in all cases
[#249] by Lachlanas.data.frame<SpatRaster> or <SpatVector>
are now also implemented as S3 methods to assure correct dispatch by
other S3 methods such as data.table::as.data.table. See
[#284] by Patrick Schratzcrs now shows the correct authority if it is not EPSG.
[#419] by Matthew Williamsonrast<filename> has a new argument
lyrs to subset the layers and open the file in one
step.rast<array> now has a crs and extent argument.
[#439] by RS-ecotype="xyz" is now default in
rast<data.frame>. [#438] by RS-ecoclassify has a new argument brackets to
show if a side of an interval is open or closed.freq and
as.data.frame. [#441] ngould7vect<matrix> and
vect<data.frame> are now much faster. [#413] by
BastienFRextract with points provided as a matrix or cell
numbers is not much faster. [#341]focal has a new argument na.policy that
can be set to one of “all” (default), “only” or “omit”. argument
na.only has been removed, as you can now use
na.policy="only"inset argument border changed to
perimeter to allow passing border on to
plot<Spat*>. [#456] by Márcia Barbosamemfrac can now be set to zero (in stead
of not lower than 0.1). [#476] by Matt Strimas-MackeyallowGaps in patches to
disallow gaps between patch IDs. See [#478] by Dunbar Carpenter.mergeTime to combine multiple rasters,
perhaps partly overlapping in time, into a single time seriesfillTime that can add empty layers in
between existing layers to assure that the time step between layers is
constantapproximate to fill in missing values by
cell across layersis.bool and as.bool for
SpatRaster and explicit recognition of Boolean raster data in various
places (e.g., extract, plot)is.int and as.int for
SpatRaster.layerCor (like
raster::layerStats). [#420] by Alex IlichfocalCor (like
raster::corLocal). [#427] by Zhuonan Wangall.equal for SpatRaster. See
[#428] by Dongdong Kongmath for SpatRaster that
implements the Math-generic methods and accepts a filenamesds<array>rasterize<matrix>, see [#413] by
BastienFRcolorize to transform color
representationsarrow to draw a (North) arrow on a map.
[#461] by Márcia Barbosadensify to insert nodes between existing
nodes of a line or polygon SpatVectordirection for SpatRaster. [#462] by Márcia
Barbosafocal3D to compute focal values for a
three-dimensional (row, column, layer) windowmakeVRT to create a vrt file for a file
that needs a header to be read.method="stratified" for
spatSample<SpatRaster>. [#470] by Michael
Mahoneymemmax to cap the amount of RAM that
terra can be used in raster processing [#476] by Matt
Strimas-MackeygridDistance to compute distances traversing
a raster, perhaps with obstacles. [#477] by Márcia BarbosaReleased 2021-11-24
focal now has ellipses (...) to allow for
providing additional arguments to fun. For this reason it
does not have a na.rm argument anymore as that can be
supplied via the ellipses. In practice this means that the default will
be na.rm=FALSE for the standard functions such as
mean and sum.app grossly overestimated RAM needed, slowing it down.
Reported by Jerry Nelsonterra now installs, again, with older versions of GEOS
[#406] by fparyaniterra did not install with Clang on CRAN/OSX due to
using C++13 idiom.lapp and tapp now have a
cores argument (as do app and
predict). Suggested by Dongdong Kong [#365]focal now also works with a function that returns
multiple values. See [#318] by Alex Ilich.focal can now process multiple layers in one step.stars objects
[#220] by Jakub NowosadfocalCpp takes a C++ function that iterates over cells
to speed up computation by avoiding apply (see [#318] by
Alex Ilich).focalReg for focal OLS regression models between
layersReleased 2021-11-16
terra did not install with versions of GDAL below 3
[#402] by Alex Ilich.distance between two SpatVectors or matrices with
pairwise=FALSE returned a matrix that was filled by column
instead of by row [#403] by Paul SmithReleased 2021-11-15
rast with some NetCDF files failed because of bad
parsing of dates. [#361] by Juan Carlos Zamora-Pereiradistance<SpatRaster> with lon/lat data was not
correct. [#368] by Greg Schmidtas.polygons<SpatRaster> failed with a SpatRaster
and a categorical layer that is not the first layer. [#370] by Patrick
Schratzrasterize was not ignored,
also causing errors when writing to temporary files. [#377] by Robbie
Pricerast<character> crashed if the sds was an empty
character string. [#381] by Dan Bastonplot<SpatVector> now responds to the
range argument [#385] by Márcia Barbosazonal failed for user-defined functions. [#393] by
mqueinnecselectHighest to select n cell values with
the highest or lowest values.vect<list> to append SpatVectors
(faster than do.call(rbind, x))align=FALSE to project to
align to the template SpatRaster but ignore the resolutiongdalCache to set the GDAL cache size,
contributed by Dan Baston [#387]fileBlocksizeoptions to writeVector to
pass layer creation options to GDALmergeLines,
snap, makeNodes, removeDupNodes,
gaps, simplifywidth and
clearanceterra now installs with older versions of GEOS
[#363]terra now installs on CentOS 7 with GDAL 2.1.4 and a
C++ compiler that does not support std::regexp. [#384] by Ariel
PaulsonReleased 2021-10-11
setValues now has two arguments
(x and values), just like raster
had; to avoid reverse dependency problems with rasterReleased 2021-10-07
To avoid name conflicts with sp (via
raster) disaggregate is now called
disagg and bbox,SpatRaster and
bbox<SpatVector> have been removed (but could be
resurrected in raster or under another name).
project and resample now choose the
resampling method based on the first layer, using “near” for categorical
data. Thanks to Matthew Lewis [#355]hist failed with small samples. Issue [#356] by Martin
QueinnecReleased 2021-10-05
terra no longer depends on raster. To avoid
name clashes between these two packages, and to allow replacing methods
from rgeos and rgdal in raster,
raster now depends on terra instead.
freq has a new argument usenames. See
issue [#309] by Bappa Dasrast<character> has a new argument
opts that can be used to pass GDAL open options. See issue
[#314]rast<SpatRaster> now takes arguments
names and vals. See issue [#323] by Dongdong
Kongcrs<- now warns if an unsupported datum is used. See
issue [#317]spatSample now returns factor values if a SpatRaster
layer is.factor except when using as.df=FALSEorigin<- to set the origin of a
SpatRaster. See issue [#326] by Jakub Nowosadcrs has a new argument parse. See
[#344]plot<SpatRaster,missing> has a new argument
reset=FALSE that allows resetting the par()$mar parameters
after plotting. See issue [#340] by Derek Friendcrds has a new argument na.rm. See [#338]
by Kodi Arfershow(Spat*) now prints the name and EPSG code of a crs
if available. See [#317] by Jakub NowosadplotRGB failed if there were NAs. Issue
[#308] by Jakub NowosadwriteVector crashed R when used with a SpatVector with
no geometries. Reported by Timothy White in issue [#319]summary<SpatRaster> now returns counts for the
classes (instead of a numerical summary of the indices) [#324] by Jakub
Nowosadtapp with a character index now returns a SpatRaster
with the correct names [#345] by Stuart Brownrasterize with a character variable now adds the ID
column to the categories [#337] by Tate BraselcellSize now masks values in all cases (when requested
with mask=TRUE). Issue [#339] by Jean-Luc Dupoueybuffer<SpatVector> no longer treats lines like
polygons [#332] by Márcia Barbosaplot now passes the layer index to fun
[#310] by Ben Tupperto_id in nearest was sometimes wrong.
See [#328] by Shawn Ligockifocal did not reset initial values for NA cells when
processing chunks. [#312] by Jeffrey Evansfocal could run out of memory when using a large window
and user-defined function, and was inexact at the chunk boundary
[#347]zonal with as.raster=TRUE failed for
categorical SpatRasters [#348] by Jakub NowosadReleased 2021-08-20
time(x) <- d is set with a Date
class object, time(x) now returns a Date
object instead of a POSIXct object. Issue [#256] by
Mauricio Zambrano-Bigiarinirast<data.frame> method to avoid confusion with
the matrix and list methods in response to a
SO question by
StackbeanswriteCDF now also writes the time dimensions if there
is only one time-step. See this SO questionvect<character> (filename) now has argument
layer to select a layer from a multi-layer file / database,
and arguments query, extent and
filter for reading a subsetsubst can now create multiple output layers See [issue
276] by Agustin Loboclassify can now create different multiple output
layers See [issue 276] by Agustin Loboalpha of plot<SpatRaster>
can now be a SpatRaster. See this SO question by
James McCarthyfilename and overwrite arguments were
ignored in rasterizeproject. Reported by
Fabian Fischerrasterize<SpatRaster,SpatVector> with
inverse=TRUE crashed the R session. Issue [#264] by
Jean-Luc Dupoueymerge and mosaic was not
correct for large rasters (only the first rows were used). Reported by
Zavud Baghirov in [#271]as.points,SpatRaster did not remove NA’s
correctly and shifted values. Issues [#269] and [#273] by Julian
Hagenauerrast<matrix> rotated values when using an
equal-sided matrix [#274] by Jakub Nowosadproject with a crs argument. [#283] by Timothée Giraudclassify, argument right had TRUE and
FALSE reversed.terrain had edge effects [#303] by Andrew Gene
Brown.terrain can now compute multiple variables at once
[#286] by Žan Kuraltwrap<SpatRaster> changed factors into numeric
[#302] by Patrick SchratzwriteVector failed with “FlatGeobuf” (and probably
other formats as well) for not using a proper MultiPolygon [#299] by L
DalbyspatSample is now
much more regular [#289] by Jakub NowosadReleased 2021-06-20
na.omit<SpatVector> to remove empty geometries
and/or attribute records that have an NAsrc to create a
SpatRasterCollection (a loose collection of tiles).merge and mosaic now have methods for a
SpatRasterCollection. To avoid the (inefficient) use of
do.call. #210 by Matthew Talluto.activeCat and activeCat<- to get or set
the “active” category if there are multiple categories (raster
attributes)as.numeric and catalyze to transfer
categories to numeric cell valuesrange and mean
for (the attributes of) a SpatVectorshade, to compute hill shadingna.rm) are now used by
rasterize with point geometries. #209 by Jakub Nowosadgstat models
by interpolate. #208 by Jakub Nowosadcpkgs to predict to list the
packages that need to be exported to the cores if argument
cores is larger than one. ?predict now shows
different approaches to parallelize predict (based on
examples in issue. #178 by by Matthew Coghill.freq now returns labels for categorical layersadjacent now has a pairs argument. #239 by
Kenneth Blake Vernonadjacent now also takes a matrix to specify adjacent
cellsmean and other summarize methods now take a
filename argument and disallow non-recognized named
arguments. #238 by Jessica Nephinvat.dbf file is now ignored if it only has the counts
of the values. #234 by Julleegdal="" write
options and use BIGTIFF if necessary (suggested by Ani Ghosh)layer in the
extract-SpatRaster,SpatVector method to extract values for
a single layers specified for each geometry (see this question).extract with points and cells=TRUE or
xy=TRUE gave garbled outputas.character<SpatRaster> (called by
wrap) did not capture the layer names. [#213] by Pascal
Titlefocal mirrored the weight matrix, thus affecting the
results when using an asymmetrical weight matrix. Reported by Sebastiano
Trevisaniterra::terraOptions now works without attaching the
package. [#229] by Karl Dunkle Wernerapp with ncores > 0 and a function that
returns multiple layers now works. [#240] by BastienFR.autocor (local) can now handle NA values.
[#245] by Jakub Nowosad .mask with a SpatVector and a large (out of memory)
multi-layer SpatRaster only worked for the first layer. Reported by
Monika Tomaszewska.Released 2021-05-13
as.lines method for SpatRasteras.polygons method for SpatVector linesautocor<numeric> has new methods
mean, to compute the local mean, and locmor,
for the local Moran’s IsharedPaths method for SpatVector (lines and
polygons)RGB2col method to reduce a three-layer RGB SpatRaster
to a single layer SpatRaster with a color-table (with <= 256
colors)split methods for SpatVector and SpatRasterrast<Raster*> now takes the crs from the Raster
object, not from the file it may point to. [#200] by Floris
Vanderhaegheconvhull has a new argument by="" to make
convex hulls for sub-sets of a SpatVector.extract with multiple layers could return a data.frame
where the values were not in the correct order (by row instead of by
column)crop works again with sf objects. [#201]
by Sebastian Brinkmannvect<sf> now also works for lines, and should be
fastervect<character> crashed R if a file had empty
geometries. [#202] by consumereextract(points, bilinear=TRUE, cells=TRUE) now works.
[#203] by fab4appzonal now works for min and
max. [#207] Reported by Jakub NowosadTo avoid name conflicts with the spatstat package
area,SpatRaster-method(x, sum=FALSE) ->
cellSize(x)area,SpatRaster/SpatVector-method(x, sum=TRUE) ->
expanse(x)convexhull -> convHullperimeter -> perimtiles -> makeTilescoords -> crdsReleased 2021-04-30
trim has a new argument value that allows
trimming rows and columns with other values than the default
NArapp has a new argument clamp that allows
clamping start and end values to 1:nlyr(x), avoiding that
all values are considered NAspatSample<SpatRaster> has new arguments
as.points and values. Getting values, cells
and coordinates is no longer mutually exclusive. In response to [#191]
by Agustin Loboarea<SpatRaster> has a new argument
mask=FALSEclassify can now take a single number to request that
many cutsmosaic and merge now warn and resample if
rasters are not alignedextract has a new argument exact to get
the fraction covered for each cellflip(x, direction="vertical") no longer reverses the
order of the layersextract did not work for horizontal or vertical lines
as their extent was considered invalid. Reported by Monika
Tomaszewskaautocor did not handle NA values [#192] by Laurence
Hawkernearest now works for angular coordinatesslope in terrain was not
correct (the tangent was returned instead of the slope), [#196] by Sven
Alderquantile now works for rasters that have cells that are
all NA. Reported by Jerry NelsonTo avoid name conflicts with tidyverse
with deprecation warning:
without deprecation warning:
Released 2021-04-14
c<SpatVector> now returns a list.
rbind is used to append SpatVector objectsrats has been removed,
and levels and cats have changed