collapse package: sum over two vectors but keep empty intersections - r

I would like to aggregate a vector/ matrix y by two variables a and b via the fsum function of the collapse package. fsum does not return values for empty intersections. Is there a way to keep empty intersection using the collapse package? I know that I could e.g. work through cross-joins and data.table, but as my function input is a vector and speed really matters, I would like to avoid converting the input matrix to a data.table and then convert the output back to a matrix / vector (for a solution with data.table, see e.g. here: data.table calculate sums by two variables and add observations for "empty" groups).
Here is an example:
library(collapse)
set.seed(1)
a <- sample(1:5, 10, replace = TRUE)
b <- sample(1:3, 10, replace = TRUE)
y <- matrix(rnorm(10), 10, 1)
fsum(x = y, g = data.frame(a = a, b = b))
#> fsum(x = y, g = data.frame(a = a, b = b))
# [,1]
#1.1 -0.40955189
#1.2 -0.05710677
#2.2 0.50360797
#2.3 -1.28459935
#3.1 0.04672617
#3.2 -0.69095384
#3.3 -0.23570656
#4.1 0.80418951
#5.2 1.08576936
What I would like to get: the regular output above, but keeping the empty intersections of (a, b) - e.g (a = 1, b = 3) and assign a missing or zero:
# a b y
#1: 1 1 -0.7702614
#2: 1 2 -0.2992151
#3: 1 3 NA
#4: 2 1 NA
#5: 2 2 -0.4115108
#6: 2 3 0.4356833
#.................
As an addition: base::aggregate() has a function argument drop = FALSE that achieves this:
aggregate(y, data.frame(a, b), sum, drop = FALSE)
a b V1
#1 1 1 -0.7702614
#2 2 1 NA
#3 3 1 -1.2375384
#4 4 1 -0.2894616
#5 5 1 NA
#6 1 2 -0.2992151
#7 2 2 -0.4115108
#8 3 2 -0.8919211
#9 4 2 NA
#10 5 2 0.2522234
#11 1 3 NA
#12 2 3 0.4356833
#13 3 3 -0.2242679
#14 4 3 NA
#15 5 3 NA
Nevertheless, in my experience both data.table and collapse are significantly faster, butcollapse has the advantage that it also works with matrix objects (that do not need to be converted to data.table's).
Is there away to achieve this via collapse?

yes you can do that with fsum, however other functions like fmedian will warn about that. To do that you need to create factors and interact them using : like so:
library(collapse)
set.seed(1)
a <- sample(1:5, 10, replace = TRUE)
b <- sample(1:3, 10, replace = TRUE)
y <- matrix(rnorm(10), 10, 1)
fsum(x = y, g = qF(a):qF(b))
# [,1]
# 1:1 -0.7702614
# 1:2 -0.2992151
# 1:3 NA
# 2:1 NA
# 2:2 -0.4115108
# 2:3 0.4356833
# 3:1 -1.2375384
# 3:2 -0.8919211
# 3:3 -0.2242679
# 4:1 -0.2894616
# 4:2 NA
# 4:3 NA
# 5:1 NA
# 5:2 0.2522234
# 5:3 NA
For the earlier example you gave, I'd also like to note that the expensive call to data.frame is absolutely not necessary, fsum(x = y, g = list(a = a, b = b)) is sufficient.

Related

How to merge elements of atomic vector in R?

I wanted to merge different elements of atomic vectors by elements names stored in list. See example:
ls = list(a = c(a = 1, b = 2, d = 2), b = c(b = 2, c = 3), c = c(a = 1, b = 2))
Now, I wanted to get output like this:
a b c
a 1 NA 1
b 2 2 2
c NA 3 NA
d 2 NA NA
I tried Reduce, but it is not working. I do not want to use any external package for this problem.
Thanks
You can use [ in sapply after you have extracted all elements names.
i <- sort(unique(unlist(lapply(ls, names))))
x <- sapply(ls, "[", i)
rownames(x) <- i
x
# a b c
#a 1 NA 1
#b 2 2 2
#c NA 3 NA
#d 2 NA NA
We could also use bind_rows here
library(dplyr)
library(tibble)
bind_rows(ls, .id = 'x') %>%
column_to_rownames('x') %>%
t
a b c
a 1 NA 1
b 2 2 2
d 2 NA NA
c NA 3 NA
Or using base R
xtabs(values ~ ind + x, do.call(rbind, Map(cbind, x = names(ls), lapply(ls, stack))))
x
ind a b c
a 1 0 1
b 2 2 2
d 2 0 0
c 0 3 0
A data.table option using rbindlist
> t(rbindlist(Map(function(x) data.table(t(x)), lst), fill = TRUE))
[,1] [,2] [,3]
a 1 NA 1
b 2 2 2
d 2 NA NA
c NA 3 NA

R: data.table left outer join function not updating

Based on this previous post I build leftOuterJoin which is a function to update a data.table X according to an other data.table Y. The function is defined as follows:
leftOuterJoin <- function(X, Y, onCol) {
.colsY <- names(Y)
X[Y, (.colsY) := mget(paste0("i.", .colsY)), on = onCol]
}
The function works 99% of the time as intended, e.g.:
X <- data.table(id = 1:5, L = letters[1:5])
id L
1: 1 a
2: 2 b
3: 3 c
4: 4 d
5: 5 e
Y <- data.table(id = 3:5, L = c(NA, "g", "h"), N = c(10, NA, 12))
id L N
1: 3 <NA> 10
2: 4 g NA
3: 5 h 12
leftOuterJoin(X, Y, "id")
X
id L N
1: 1 a NA
2: 2 b NA
3: 3 <NA> 10
4: 4 g NA
5: 5 h 12
However, for some reason that is unknown to me, it just stops working with some data tables (I have no reproductible example at hand). There is no error, but the data table is not updated. When I use the debug function, everything seems to be working fine, X is updated, but the real data.table isn't. Now, if I just do it outside the function it works. Maybe it is related to the scope of the function? I am really struggling with this problem.
Spec: R v3.5.1 and data.table v1.11.4.
EDIT
Based on the comments I figured out that the problem is related to the data.table pointer. You can reproduce the problem with this code:
> save(X, file = "X.RData")
> load("X.RData")
> leftOuterJoin(X, Y, "id")
> X
id L
1: 1 a
2: 2 b
3: 3 <NA>
4: 4 g
5: 5 h
Notice that X is updated but not the way we want it. However, if we use setDT() it works properly:
> load("X.RData")
> setDT(X)
> leftOuterJoin(X, Y, "id")
> X
id L N
1: 1 a NA
2: 2 b NA
3: 3 <NA> 10
4: 4 g NA
5: 5 h 12
Is there a way to set up leftOuterJoin() such that it will not be necessary to run setDT() every time some data is loaded?

remove cases following certain other cases

I have a dataframe, say
df = data.frame(x = c("a","a","b","b","b","c","d","t","c","b","t","c","t","a","a","b","d","t","t","c"),
y = c(2,4,5,2,6,2,4,5,2,6,2,4,5,2,6,2,4,5,2,6))
I want to remove only those rows in which one or multiple ts are directly in between a d and a c, in all other cases I want to retain the cases. So for this example, I would like to remove the ts on row 8, 18 and 19, but keep the others. I have over thousands of cases so doing this manually would be a true horror. Any help is very much appreciated.
One option would be to use rle to get runs of the same string and then you can use an sapply to check forward/backward and return all the positions you want to drop:
rle_vals <- rle(as.character(df$x))
drop <- unlist(sapply(2:length(rle_vals$values), #loop over values
function(i, vals, lengths) {
if(vals[i] == "t" & vals[i-1] == "d" & vals[i+1] == "c"){#Check if value is "t", previous is "d" and next is "c"
(sum(lengths[1:i-1]) + 1):sum(lengths[1:i]) #Get row #s
}
},vals = rle_vals$values, lengths = rle_vals$lengths))
drop
#[1] 8 18 19
df[-drop,]
# x y
#1 a 2
#2 a 4
#3 b 5
#4 b 2
#5 b 6
#6 c 2
#7 d 4
#9 c 2
#10 b 6
#11 t 2
#12 c 4
#13 t 5
#14 a 2
#15 a 6
#16 b 2
#17 d 4
#20 c 6
This also works, by collapsing to a string, identifying groups of t's between d and c (or c and d - not sure whether you wanted this option as well), then working out where they are and removing the rows as appropriate.
df = data.frame(x=c("a","a","b","b","b","c","d","t","c","b","t","c","t","a","a","b","d","t","t","c"),
y=c(2,4,5,2,6,2,4,5,2,6,2,4,5,2,6,2,4,5,2,6),stringsAsFactors = FALSE)
dfs <- paste0(df$x,collapse="") #collapse to a string
dfs2 <- do.call(rbind,lapply(list(gregexpr("dt+c",dfs),gregexpr("ct+d",dfs)),
function(L) data.frame(x=L[[1]],y=attr(L[[1]],"match.length"))))
dfs2 <- dfs2[dfs2$x>0,] #remove any -1 values (if string not found)
drop <- unlist(mapply(function(a,b) (a+1):(a+b-2),dfs2$x,dfs2$y))
df2 <- df[-drop,]
Here is another solution with base R:
df = data.frame(x = c("a","a","b","b","b","c","d","t","c","b","t","c","t","a","a","b","d","t","t","c"),
y = c(2,4,5,2,6,2,4,5,2,6,2,4,5,2,6,2,4,5,2,6))
#
s <- paste0(df$x, collapse="")
L <- c(NA, NA)
while (TRUE) {
r <- regexec("dt+c", s)[[1]]
if (r[1]==-1) break
L <- rbind(L, c(pos=r[1]+1, length=attr(r, "match.length")-2))
s <- sub("d(t+)c", "x\\1x", s)
}
L <- L[-1,]
drop <- unlist(apply(L,1, function(x) seq(from=x[1], len=x[2])))
df[-drop, ]
# > drop
# 8 18 19
# > df[-drop, ]
# x y
# 1 a 2
# 2 a 4
# 3 b 5
# 4 b 2
# 5 b 6
# 6 c 2
# 7 d 4
# 9 c 2
# 10 b 6
# 11 t 2
# 12 c 4
# 13 t 5
# 14 a 2
# 15 a 6
# 16 b 2
# 17 d 4
# 20 c 6
With gregexpr() it is shorter:
s <- paste0(df$x, collapse="")
g <- gregexpr("dt+c", s)[[1]]
L <- data.frame(pos=g+1, length=attr(g, "match.length")-2)
drop <- unlist(apply(L,1, function(x) seq(from=x[1], len=x[2])))
df[-drop, ]

Calculate cumsum() while ignoring NA values

Consider the following named vector x.
( x <- setNames(c(1, 2, 0, NA, 4, NA, NA, 6), letters[1:8]) )
# a b c d e f g h
# 1 2 0 NA 4 NA NA 6
I'd like to calculate the cumulative sum of x while ignoring the NA values. Many R functions have an argument na.rm which removes NA elements prior to calculations. cumsum() is not one of them, which makes this operation a bit tricky.
I can do it this way.
y <- setNames(numeric(length(x)), names(x))
z <- cumsum(na.omit(x))
y[names(y) %in% names(z)] <- z
y[!names(y) %in% names(z)] <- x[is.na(x)]
y
# a b c d e f g h
# 1 3 3 NA 7 NA NA 13
But this seems excessive, and makes a lot of new assignments/copies. I'm sure there's a better way.
What better methods are there to return the cumulative sum while effectively ignoring NA values?
You can do this in one line with:
cumsum(ifelse(is.na(x), 0, x)) + x*0
# a b c d e f g h
# 1 3 3 NA 7 NA NA 13
Or, similarly:
library(dplyr)
cumsum(coalesce(x, 0)) + x*0
# a b c d e f g h
# 1 3 3 NA 7 NA NA 13
It's an old question but tidyr gives a new solution.
Based on the idea of replacing NA with zero.
require(tidyr)
cumsum(replace_na(x, 0))
a b c d e f g h
1 3 3 3 7 7 7 13
Do you want something like this:
x2 <- x
x2[!is.na(x)] <- cumsum(x2[!is.na(x)])
x2
[edit] Alternatively, as suggested by a comment above, you can change NA's to 0's -
miss <- is.na(x)
x[miss] <- 0
cs <- cumsum(x)
cs[miss] <- NA
# cs is the requested cumsum
Here's a function I came up from the answers to this question. Thought I'd share it, since it seems to work well so far. It calculates the cumulative FUNC of x while ignoring NA. FUNC can be any one of sum(), prod(), min(), or max(), and x is a numeric vector.
cumSkipNA <- function(x, FUNC)
{
d <- deparse(substitute(FUNC))
funs <- c("max", "min", "prod", "sum")
stopifnot(is.vector(x), is.numeric(x), d %in% funs)
FUNC <- match.fun(paste0("cum", d))
x[!is.na(x)] <- FUNC(x[!is.na(x)])
x
}
set.seed(1)
x <- sample(15, 10, TRUE)
x[c(2,7,5)] <- NA
x
# [1] 4 NA 9 14 NA 14 NA 10 10 1
cumSkipNA(x, sum)
# [1] 4 NA 13 27 NA 41 NA 51 61 62
cumSkipNA(x, prod)
# [1] 4 NA 36 504 NA 7056 NA
# [8] 70560 705600 705600
cumSkipNA(x, min)
# [1] 4 NA 4 4 NA 4 NA 4 4 1
cumSkipNA(x, max)
# [1] 4 NA 9 14 NA 14 NA 14 14 14
Definitely nothing new, but maybe useful to someone.
Another option is using the collapse package with fcumsum function like this:
( x <- setNames(c(1, 2, 0, NA, 4, NA, NA, 6), letters[1:8]) )
#> a b c d e f g h
#> 1 2 0 NA 4 NA NA 6
library(collapse)
fcumsum(x)
#> a b c d e f g h
#> 1 3 3 NA 7 NA NA 13
Created on 2022-08-24 with reprex v2.0.2

How to squeeze in missing values into a vector

Let me try to make this question as general as possible.
Let's say I have two variables a and b.
a <- as.integer(runif(20, min = 0, max = 10))
a <- as.data.frame(a)
b <- as.data.frame(a[c(-7, -11, -15),])
So b has 17 observations and is a subset of a which has 20 observations.
My question is the following: how I would use these two variables to generate a third variable c which like a has 20 observations but for which observations 7, 11 and 15 are missing, and for which the other observations are identical to b but in the order of a?
Or to put it somewhat differently: how could I squeeze in these missing observations into variable b at locations 7, 11 and 15?
It seems pretty straightforward (and it probably is) but I have been not getting this to work for a bit too long now.
1) loop Try this loop:
# test data
set.seed(123) # for reproducibility
a <- as.integer(runif(20, min = 0, max = 10))
a <- as.data.frame(a)
b <- as.data.frame(a[c(-7, -11, -15),])
# lets work with vectors
A <- a[[1]]
B <- b[[1]]
j <- 1
C <- A
for(i in seq_along(A)) if (A[i] == B[j]) j <- j+1 else C[i] <- NA
which gives:
> C
[1] 2 7 4 8 9 0 NA 8 5 4 NA 4 6 5 NA 8 2 0 3 9
2) Reduce Here is a loop-free version:
f <- function(j, a) j + (a == B[j])
r <- Reduce(f, A, acc = TRUE)
ifelse(duplicated(r), NA, A)
giving:
[1] 2 7 4 8 9 0 NA 8 5 4 NA 4 6 5 NA 8 2 0 3 9
3) dtw. Using dtw in the package of the same name we can get a compact loop-free one-liner:
library(dtw)
ifelse(duplicated(dtw(A, B)$index2), NA, A)
giving:
[1] 2 7 4 8 9 0 NA 8 5 4 NA 4 6 5 NA 8 2 0 3 9
REVISED Added additional solutions.
Here's a more complicated way of doing it, using the Levenshtein distance algorithm, that does a better job on more complicated examples (it also seemed faster in a couple of larger tests I tried):
# using same data as G. Grothendieck:
set.seed(123) # for reproducibility
a <- as.integer(runif(20, min = 0, max = 10))
a <- as.data.frame(a)
b <- as.data.frame(a[c(-7, -11, -15),])
A = a[[1]]
B = b[[1]]
# compute the transformation between the two, assigning infinite weight to
# insertion and substitution
# using +1 here because the integers fed to intToUtf8 have to be larger than 0
# could also adjust the range more dynamically based on A and B
transf = attr(adist(intToUtf8(A+1), intToUtf8(B+1),
costs = c(Inf,1,Inf), counts = TRUE), 'trafos')
C = A
C[substring(transf, 1:nchar(transf), 1:nchar(transf)) == "D"] <- NA
#[1] 2 7 4 8 9 0 NA 8 5 4 NA 4 6 5 NA 8 2 0 3 9
More complex matching example (where the greedy algorithm would perform poorly):
A = c(1,1,2,2,1,1,1,2,2,2)
B = c(1,1,1,2,2,2)
transf = attr(adist(intToUtf8(A), intToUtf8(B),
costs = c(Inf,1,Inf), counts = TRUE), 'trafos')
C = A
C[substring(transf, 1:nchar(transf), 1:nchar(transf)) == "D"] <- NA
#[1] NA NA NA NA 1 1 1 2 2 2
# the greedy algorithm would return this instead:
#[1] 1 1 NA NA 1 NA NA 2 2 2
The data frame version, which isn't terribly different from G.'s above.
(Assumes a,b setup as above).
j <- 1
c <- a
for (i in (seq_along(a[,1]))) {
if (a[i,1]==b[j,1]) {
j <- j+1
} else
{
c[i,1] <- NA
}
}

Resources