Subsetting neighboring fileds

0
votes

I am trying to make a conditional subsetting which would include contemporary elements inside window of neighboring areas. For example, given the matrix Dat, where Species (SP), Area (AR) and Time (TM):

SP AR TM
A  2  2
B  2  2
C  1  4
F  3  2
B  5  3
E  3  2
D  2  1
I  1  4
H  3  2
E  2  4
D  3  5
B  1  2

How can I retrieve all the species co-occurring with species A in the same time, but within neighboring areas (in this case 1 and 3)? The desired output being:

SP  AR  TM
A  2  2
B  2  2
F  3  2
H  3  2
B  1  2

This is based on the assumption that Species A will be occurring repeatedly in the dataset in different areas. I have an attempt, given by user thelatemail from a different question (with related elements) I posted previously, with minor modifications. The X's added, indicate the part of the syntax I cannot figure out, which would basically be the definition of the bracket (given this is more or less where it should be going).

with(dat,dat[
  apply(
    sapply(TM[SP=="A"],
    function(x) abs(AR)XXXXX),1,any
  )
,]
)

Any help is much appreciated.

This is part of a set of operations I am trying to make on a large dataset. I broke the question into two elements, which are somewhat related but far from being duplicates. The reason for this being I am a beginning R user and want to learn how to interpret, write and integrate codes on my own. The link to the related question: Subsetting based on co-occurrence within a time window . I can remove one of the links if needed.

1
rsubsetintervalssapply
Subsetting in R is well covered in many tutorials on line. dat[dat$TM==2,] gives your desired result... however more specifically: dat[dat$TM==2 & dat$AR %in% 1:3,] where that 1:3 can be a separate function to define "neighboring" areas. - Justin
Generally, it is good form to at least link to the critical question that this is derrived from. As it stands, one must go research your earlier questions and parse the solution there. Without that knowledge, the answer in my comment is correct and much simpler. - Justin

1 Answers

1
votes

Assuming neighboring areas are defined by areas within +/- 1 of the given area:

Copying from the previous question: Subsetting based on co-occurrence within a time window ...

with(dat,dat[
  (
    SP=="A" |
    # Area %in% Area[SP=="A"]
    Area %in% c(Area[SP=='A']-1, Area[SP=='A'], Area[SP=='A']+1)
  ) & 
  apply(
    sapply(Time[SP=="A"],
    function(x) abs(difftime(Time,x,units="mins"))<=30 ),1,any
  ) 
,]
)

As per the comment (and my comment)

area <- 2
dat[dat$AR %in% c(area - 1, area, area + 1),]

And in Regards to Conditional subsetting by POSIXct interval and another field containing interval removing the conditional for SP=='A' should result in correct subsetting

area_boolean <- with(dat, Area %in% c(Area[SP=='A']-1, Area[SP=='A'], Area[SP=='A']+1))
time_boolean <- with(dat, apply(sapply(Time[SP=="A"],
                                function(x) abs(difftime(Time, x, units="mins")) <= 30 ),  
                                1, 
                                any))
dat[area_boolean & time_boolean,]