# Extract.ppp

##### Extract or Replace Subset of Point Pattern

Extract or replace a subset of a point pattern. Extraction of a subset has the effect of thinning the points and/or trimming the window.

##### Usage

```
## S3 method for class 'ppp':
[(x, i, j, drop, ...)
## S3 method for class 'ppp':
[(x, i, j) <- value
```

##### Arguments

- x
- A two-dimensional point pattern.
An object of class
`"ppp"`

. - i
- Subset index. Either a valid subset index in the usual Rsense,
indicating which points should be retained, or a window
(an object of class
`"owin"`

) delineating a subset of the original observation window. - value
- Replacement value for the subset. A point pattern.
- j
- Redundant. Included for backward compatibility.
- drop, ...
- Ignored. These arguments are required for compatibility with the generic function.

##### Details

These functions extract a designated subset of a point pattern, or replace the designated subset with another point pattern.

The function `[.ppp`

is a method for `[`

for the
class `"ppp"`

. It extracts a designated subset of a point pattern,
either by ``*thinning*''
(retaining/deleting some points of a point pattern)
or ``*trimming*'' (reducing the window of observation
to a smaller subregion and retaining only
those points which lie in the subregion) or both.

The pattern will be ``thinned''
if `i`

is a subset index in the usual Rsense:
either a numeric vector
of positive indices (identifying the points to be retained),
a numeric vector of negative indices (identifying the points
to be deleted) or a logical vector of length equal to the number of
points in the point pattern `x`

. In the latter case,
the points `(x$x[i], x$y[i])`

for which
`subset[i]=TRUE`

will be retained, and the others
will be deleted.
The pattern will be ``trimmed''
if `i`

is an object of class
`"owin"`

specifying a window of observation.
The points of `x`

lying inside the new
`window`

will be retained. Alternatively `i`

may be a
pixel image (object of class `"im"`

) with logical values;
the pixels with the value `TRUE`

will be interpreted as a window.
The function `[<-.ppp`

is a method for `[<-`

for the
class `"ppp"`

. It replaces the designated
subset with the point pattern `value`

.
The subset of `x`

to be replaced is designated by
the argument `i`

as above.

The replacement point pattern `value`

must lie inside the
window of the original pattern `x`

.
The ordering of points in `x`

will be preserved
if the replacement pattern `value`

has the same number of points
as the subset to be replaced. Otherwise the ordering is
unpredictable.

If the original pattern `x`

has marks, then the replacement
pattern `value`

must also have marks, of the same type.

Use the function `unmark`

to remove marks from a
marked point pattern.

Use the function `split.ppp`

to select those points
in a marked point pattern which have a specified mark.

##### Value

- A point pattern (of class
`"ppp"`

).

##### Warnings

The function does not check whether `window`

is a subset of
`x$window`

. Nor does it check whether `value`

lies
inside `x$window`

.

##### See Also

##### Examples

```
data(longleaf)
# Longleaf pines data
plot(longleaf)
<testonly>longleaf <- longleaf[seq(1,longleaf$n,by=10)]</testonly>
# adult trees defined to have diameter at least 30 cm
adult <- (longleaf$marks >= 30)
longadult <- longleaf[adult]
plot(longadult)
# note that the marks are still retained.
# Use unmark(longadult) to remove the marks
# New Zealand trees data
data(nztrees)
plot(nztrees) # plot shows a line of trees at the far right
abline(v=148, lty=2) # cut along this line
nzw <- owin(c(0,148),c(0,95)) # the subwindow
# trim dataset to this subwindow
nzsub <- nztrees[nzw]
plot(nzsub)
# Redwood data
data(redwood)
plot(redwood)
# Random thinning: delete 60\% of data
retain <- (runif(redwood$n) < 0.4)
thinred <- redwood[retain]
plot(thinred)
# Scramble 60\% of data
modif <- (runif(redwood$n) < 0.6)
scramble <- function(x) { runifpoint(x$n, x$window) }
redwood[modif] <- scramble(redwood[modif])
# Lansing woods data - multitype points
data(lansing)
<testonly>lansing <- lansing[seq(1, lansing$n, length=100)]</testonly>
# Hickory trees
hicks <- split(lansing)$hickory
# Trees in subwindow
win <- owin(c(0.3, 0.6),c(0.2, 0.5))
lsub <- lansing[win]
# Scramble the locations of trees in subwindow, retaining their marks
lansing[win] <- scramble(lsub) %mark% (lsub$marks)
```

*Documentation reproduced from package spatstat, version 1.27-0, License: GPL (>= 2)*