CODEWITHSUNDEEP
r
Kaja
Kaja

Reputation: 3057

need help to understand a function in r

I am trying to understand a function in R. could you please declare some part of it for me:

the function is:

subsignals <- lapply(c(peakind$freqindex, midindex+1), function(x){
upperind <- x
fsub <- f
notnullind <- ((fsub$freqindex >= lowerind
                & fsub$freqindex < upperind)
               |
                (fsub$freqindex >  (lindex - upperind + 2)
                 & fsub$freqindex <= (lindex - lowerind + 2)))
fsub[!notnullind,"coef"] <- 0
lowerind <<- upperind
Re(fft(fsub$coef, inverse=TRUE)/length(fsub$coef))
})

Could some one explain me:

1-What could be the content of notnullind and generally, what does this part of code do:

notnullind <- ((fsub$freqindex >= lowerind
                & fsub$freqindex < upperind)
               |
                (fsub$freqindex >  (lindex - upperind + 2)
                 & fsub$freqindex <= (lindex - lowerind + 2)))

2-What does fsub[!notnullind,"coef"] <- 0 mean?

3-What does<<- in lowerind <<- upperind mean?

further information: peakind looks like this:

               coef freqindex
9   2.714391+3.327237i         9
17  1.273340+4.023808i        17
25 -0.445424+5.674848i        25
33 -1.378107+3.182281i        33
41 -2.798383+2.340895i        41
49 -4.479888+1.095193i        49

and fsub :

                     coef freqindex
1   19.2352397+0.0000000i         1
2   -0.4799684+0.1651822i         2
3    1.5235726+0.0790459i         3
4   -0.1165587+0.1217513i         4
5    2.2376900+1.6763410i         5
6    1.1256711+0.4624819i         6
.....
102 -0.1165587-0.1217513i       102
103  1.5235726-0.0790459i       103
104 -0.4799684-0.1651822i       104

Upvotes: 0

Views: 89

Answers (1)

BrodieG
BrodieG

Reputation: 52687

It seems that the code is iterating through fsub in chunks defined by the difference between adjacent entries in peakind. Presumably peakind contains interesting points in fsub. You can see this because most of the fsub comparison are between x (which comes from peakind), and lowerind, which is set to be the prior loops x/upperind value.

  1. notnullind will be a logical vector (TRUE, FALSE) that is TRUE for the rows in fsub that are between this iterations peakind$freqindex and the prior ones as well as something else based on lindex that I can't tell you b/c that variable is undefined in your code.
  2. That line is setting all the values in fsub$coef that don't meet the condition described above to zero
  3. lowerind<<-upperind is a global assignment outside of the function being run through lapply. This allows the function run by lapply to keep track of the last upperind from a previous call to that same function in lapply loop. The assignment must be global as otherwise the value would be lost after every iteration in lapply.

Basically, the function is doing an fft for the data between in fsub between adjacent pairs of index values defined in peakind.

Note that 3. suggests your function isn't structured in the best possible manner. You should generally avoid global assignments unless you really can't. In this case, I would loop through the rows of cbind(head(peakind$freqindex, -1L), tail(peakind$freqindex, -1L)) which contains the range of indices you care about for each iteration.

Upvotes: 2

Related Questions