Consultation summary 2023-08-28

task 1: drop any groups with less than 10 rows

# make fake data; notice group 4 has only 8 rows
grp <- c(rep(0:3, each = 10), rep(4,8), rep(5:6, each = 10))
y <- round(rnorm(length(grp)),2)
d <- data.frame(y, grp)

Want to drop group 4 since it only has 8 rows instead of 10

table(d$grp)

 0  1  2  3  4  5  6 
10 10 10 10  8 10 10 

Here’s how I did it.

# calculate table
tab <- table(d$grp)
# save names of table with entries not equal to 10
drop <- names(tab)[tab != 10]
# subset data frame where grp is not in drop
d2 <- subset(d, !(grp %in% drop))
# group 4 is dropped
table(d2$grp)

 0  1  2  3  5  6 
10 10 10 10 10 10 

task 2: drop rows with runs of zeroes greater than or equal to 5

# make fake data with lots of zeroes
set.seed(666)
y <- sample(0:4, size = 1000, replace = TRUE, prob = c(16,1,1,1,1)/20)
d <- data.frame(y)
head(d, n = 20)

   y
1  0
2  0
3  1
4  0
5  0
6  0
7  1
8  0
9  0
10 0
11 0
12 0
13 0
14 0
15 0
16 2
17 0
18 3
19 0
20 0

Notice we start off with a run of 2 zeroes, then 1 one, then 3 zeroes, then 1 one, then 8 zeroes, …

We can use the rle() function to calculate runs.

r_out <- rle(d$y)
r_out

Run Length Encoding
  lengths: int [1:359] 2 1 3 1 8 1 1 1 2 1 ...
  values : int [1:359] 0 1 0 1 0 2 0 3 0 1 ...

lengths measures the runs of the values:

Notice we start off with a run of 2 zeroes, then 1 one, then 3 zeroes, then 1 one, then 8 zeroes, …

Again, the task is to drop rows where the run of zeroes is 5 or more. Here’s how I did it. Add the lengths to the data frame, repeating by length so each row will be identified by its run group.

d$cnt <- rep(r_out$lengths, r_out$lengths)
head(d, n = 20)

   y cnt
1  0   2
2  0   2
3  1   1
4  0   3
5  0   3
6  0   3
7  1   1
8  0   8
9  0   8
10 0   8
11 0   8
12 0   8
13 0   8
14 0   8
15 0   8
16 2   1
17 0   1
18 3   1
19 0   2
20 0   2

Now we can subset the data frame on the condition that cnt < 5.

d2 <- subset(d, cnt < 5)
r_out2 <- rle(d2$y)
any(r_out2$lengths >= 5)

[1] FALSE