Fill Missing Values - r

data=data.frame("student"=c(1,1,1,1,2,2,2,2,3,3,3,3,4),
"timeHAVE"=c(1,4,7,10,2,5,NA,11,6,NA,NA,NA,3),
"timeWANT"=c(1,4,7,10,2,5,8,11,6,9,12,15,3))
library(dplyr);library(tidyverse)
data$timeWANTattempt=data$timeHAVE
data <- data %>%
group_by(student) %>%
fill(timeWANTattempt)+3
I have 'timeHAVE' and I want to replace missing times with the previous time +3. I show my dplyr attempt but it does not work. I seek a data.table solution. Thank you.


you can try.
data %>%
group_by(student) %>%
mutate(n_na = cumsum(is.na(timeHAVE))) %>%
mutate(timeHAVE = ifelse(is.na(timeHAVE), timeHAVE[n_na == 0 & lead(n_na) == 1] + 3*n_na, timeHAVE))
student timeHAVE timeWANT n_na
<dbl> <dbl> <dbl> <int>
1 1 1 1 0
2 1 4 4 0
3 1 7 7 0
4 1 10 10 0
5 2 2 2 0
6 2 5 5 0
7 2 8 8 1
8 2 11 11 1
9 3 6 6 0
10 3 9 9 1
11 3 12 12 2
12 3 15 15 3
13 4 3 3 0
I included the little helper n_na which counts NA's in a row. Then the second mutate muliplies the number of NAs with three and adds this to the first non-NA element before NA's


Here's an approach using 'locf' filling
setDT(data)
data[ , by = student, timeWANT := {
# carry previous observations forward whenever missing
locf_fill = nafill(timeHAVE, 'locf')
# every next NA, the amount shifted goes up by another 3
na_shift = cumsum(idx <- is.na(timeHAVE))
# add the shift, but only where the original data was missing
locf_fill[idx] = locf_fill[idx] + 3*na_shift[idx]
# return the full vector
locf_fill
}]
Warning that this won't work if a given student can have more than one non-consecutive set of NA values in timeHAVE


Another data.table option without grouping:
setDT(data)[, w := fifelse(is.na(timeHAVE) & student==shift(student),
nafill(timeHAVE, "locf") + 3L * rowid(rleid(timeHAVE)),
timeHAVE)]
output:
student timeHAVE timeWANT w
1: 1 1 1 1
2: 1 4 4 4
3: 1 7 7 7
4: 1 10 10 10
5: 2 2 2 2
6: 2 5 5 5
7: 2 NA 8 8
8: 2 11 11 11
9: 3 6 6 6
10: 3 NA 9 9
11: 3 NA 12 12
12: 3 NA 15 15
13: 4 NA NA NA
14: 4 3 3 3
data with student=4 having NA for the first timeHAVE:
data = data.frame("student"=c(1,1,1,1,2,2,2,2,3,3,3,3,4,4),
"timeHAVE"=c(1,4,7,10,2,5,NA,11,6,NA,NA,NA,NA,3),
"timeWANT"=c(1,4,7,10,2,5,8,11,6,9,12,15,NA,3))

Related

Removing groups with all NA in Data.Table or DPLYR in R

dataHAVE = data.frame("student"=c(1,1,1,2,2,2,3,3,3,4,4,4,5,5,5),
"time"=c(1,2,3,1,2,3,1,2,3,NA,NA,NA,NA,2,3),
"score"=c(7,9,5,NA,NA,NA,NA,3,9,NA,NA,NA,7,NA,5))
dataWANT=data.frame("student"=c(1,1,1,3,3,3,5,5,5),
"time"=c(1,2,3,1,2,3,NA,2,3),
"score"=c(7,9,5,NA,3,9,7,NA,5))
I have a tall dataframe and in that data frame I want to remove student IDS that contain NA for all 'score' or for all 'time'. This is just if it is all NA, if there are some NA then I want to keep all their records...

Is this what you want?
library(dplyr)
dataHAVE %>%
group_by(student) %>%
filter(!all(is.na(score)))
student time score
<dbl> <dbl> <dbl>
1 1 1 7
2 1 2 9
3 1 3 5
4 3 1 NA
5 3 2 3
6 3 3 9
7 5 NA 7
8 5 2 NA
9 5 3 5
Each student is only kept if not (!) all score values are NA

Since nobody suggested one, here is a solution using data.table:
library(data.table)
dataHAVE = data.table("student"=c(1,1,1,2,2,2,3,3,3,4,4,4,5,5,5),
"time"=c(1,2,3,1,2,3,1,2,3,NA,NA,NA,NA,2,3),
"score"=c(7,9,5,NA,NA,NA,NA,3,9,NA,NA,NA,7,NA,5))
Edit:
Previous but wrong code:
dataHAVE[, .SD[!(all(is.na(time)) & all(is.na(score)))], by = student]
New and correct code:
dataHAVE[, .SD[!(all(is.na(time)) | all(is.na(score)))], by = student]
Returns:
student time score
1: 1 1 7
2: 1 2 9
3: 1 3 5
4: 3 1 NA
5: 3 2 3
6: 3 3 9
7: 5 NA 7
8: 5 2 NA
9: 5 3 5
Edit:
Updatet data.table solution with #Cole s suggestion...

Here is a base R solution using subset + ave
dataWANT <- subset(dataHAVE,!(ave(time,student,FUN = function(v) all(is.na(v))) | ave(score,student,FUN = function(v) all(is.na(v)))))
or
dataWANT <- subset(dataHAVE,
!Reduce(`|`,Map(function(x) ave(get(x),student,FUN = function(v) all(is.na(v))), c("time","score"))))

Another option:
library(data.table)
setDT(dataHAVE, key="student")
dataHAVE[!student %in% dataHAVE[, if(any(colSums(is.na(.SD))==.N)) student, student]$V1]

Create a dummy variable, and filter based on that
library("dplyr")
dataHAVE = data.frame("student"=c(1,1,1,2,2,2,3,3,3,4,4,4,5,5,5),
"time"=c(1,2,3,1,2,3,1,2,3,NA,NA,NA,NA,2,3),
"score"=c(7,9,5,NA,NA,NA,NA,3,9,NA,NA,NA,7,NA,5))
dataHAVE %>%
mutate(check=is.na(time)&is.na(score)) %>%
filter(check == FALSE) %>%
select(-check)
#> student time score
#> 1 1 1 7
#> 2 1 2 9
#> 3 1 3 5
#> 4 2 1 NA
#> 5 2 2 NA
#> 6 2 3 NA
#> 7 3 1 NA
#> 8 3 2 3
#> 9 3 3 9
#> 10 5 NA 7
#> 11 5 2 NA
#> 12 5 3 5
Created on 2020-02-21 by the reprex package (v0.3.0)

data.table solution generalising to any number of columns:
dataHAVE[,
.SD[do.call("+", lapply(.SD, function(x) any(!is.na(x)))) == ncol(.SD)],
by = student]
# student time score
# 1: 1 1 7
# 2: 1 2 9
# 3: 1 3 5
# 4: 3 1 NA
# 5: 3 2 3
# 6: 3 3 9
# 7: 5 NA 7
# 8: 5 2 NA
# 9: 5 3 5

Row conditional column operations in data.table

I have a large data.table where I for each row need to make computations based on part of the full data.table. As an example consider the following data.table, and assume I for each row want to compute the sum of the num variable for every rows where id2 matches id1 for the current row as well as the time variable is within distance 1 from the time of the current row.
set.seed(123)
dat <- data.table(cbind(id1=sample(1:5,10,replace=T),
id2=sample(1:5,10,replace=T),
num=sample(1:10,10,replace=T),
time=sample(1:10,10,replace=T)))
This could easily be done by looping over each row like this
dat[,val:= 0]
for (i in 1:nrow(dat)){
this.val <- dat[ (id2==id1[i]) & (time>=time[i]-2) & (time<=time[i]+2),sum(num)]
dat[i,val:=this.val]
}
dat
The resulting data.table looks like this:
> dat
id1 id2 num time val
1: 2 5 9 10 6
2: 4 3 7 10 0
3: 3 4 7 7 10
4: 5 3 10 8 9
5: 5 1 7 1 2
6: 1 5 8 5 6
7: 3 2 6 8 17
8: 5 1 6 3 10
9: 3 2 3 4 0
10: 3 5 2 3 0
What is the proper/fast way to do things like this using data.table?

We can use a self-join here by creating the 'timeminus2' and 'timeplus2' column, join on by 'id2' with 'id1' and the non-equi logical condition to get the sum of 'num' and assign (:=) the 'val' column to the original dataset
tmp <- dat[.(id1 = id1, timeminus2 = time - 2, timeplus2 = time + 2),
.(val = sum(num)),
on = .(id2 = id1, time >= timeminus2, time <= timeplus2),
by = .EACHI
][is.na(val), val := 0][]
dat[, val := tmp$val][]
# id1 id2 num time val
# 1: 2 5 9 10 6
# 2: 4 3 7 10 0
# 3: 3 4 7 7 10
# 4: 5 3 10 8 9
# 5: 5 1 7 1 2
# 6: 1 5 8 5 6
# 7: 3 2 6 8 17
# 8: 5 1 6 3 10
# 9: 3 2 3 4 0
#10: 3 5 2 3 0

How do I select rows in a data frame before and after a condition is met?

I'm searching the web for a few a days now and I can't find a solution to my (probably easy to solve) problem.
I have huge data frames with 4 variables and over a million observations each. Now I want to select 100 rows before, all rows while and 1000 rows after a specific condition is met and fill the rest with NA's. I tried it with a for loop and if/ifelse but it doesn't work so far. I think it shouldn't be a big thing, but in the moment I just don't get the hang of it.
I create the data using:
foo<-data.frame(t = 1:15, a = sample(1:15), b = c(1,1,1,1,1,4,4,4,4,1,1,1,1,1,1), c = sample(1:15))
My Data looks like this:
ID t a b c
1 1 4 1 7
2 2 7 1 10
3 3 10 1 6
4 4 2 1 4
5 5 13 1 9
6 6 15 4 3
7 7 8 4 15
8 8 3 4 1
9 9 9 4 2
10 10 14 1 8
11 11 5 1 11
12 12 11 1 13
13 13 12 1 5
14 14 6 1 14
15 15 1 1 12
What I want is to pick the value of a (in this example) 2 rows before, all rows while and 3 rows after the value of b is >1 and fill the rest with NA's. [Because this is just an example I guess you can imagine that after these 15 rows there are more rows with the value for b changing from 1 to 4 several times (I did not post it, so I won't spam the question with unnecessary data).]
So I want to get something like:
ID t a b c d
1 1 4 1 7 NA
2 2 7 1 10 NA
3 3 10 1 6 NA
4 4 2 1 4 2
5 5 13 1 9 13
6 6 15 4 3 15
7 7 8 4 15 8
8 8 3 4 1 3
9 9 9 4 2 9
10 10 14 1 8 14
11 11 5 1 11 5
12 12 11 1 13 11
13 13 12 1 5 NA
14 14 6 1 14 NA
15 15 1 1 12 NA
I'm thankful for any help.
Thank you.
Best regards,
Chris

here is the same attempt as missuse, but with data.table:
library(data.table)
foo<-data.frame(t = 1:11, a = sample(1:11), b = c(1,1,1,4,4,4,4,1,1,1,1), c = sample(1:11))
DT <- setDT(foo)
DT[ unique(c(DT[,.I[b>1] ],DT[,.I[b>1]+3 ],DT[,.I[b>1]-2 ])), d := a]
t a b c d
1: 1 10 1 2 NA
2: 2 6 1 10 6
3: 3 5 1 7 5
4: 4 11 4 4 11
5: 5 4 4 9 4
6: 6 8 4 5 8
7: 7 2 4 8 2
8: 8 3 1 3 3
9: 9 7 1 6 7
10: 10 9 1 1 9
11: 11 1 1 11 NA
Here
unique(c(DT[,.I[b>1] ],DT[,.I[b>1]+3 ],DT[,.I[b>1]-2 ]))
gives you your desired indixes : the unique indices of the line for your condition, the same indices+3 and -2.

Here is an attempt.
Get indexes that satisfy the condition b > 1
z <- which(foo$b > 1)
get indexes for (z - 2) : (z + 3)
ind <- unique(unlist(lapply(z, function(x){
g <- pmax(x - 2, 1) #if x - 2 is negative
g : (x + 3)
})))
create d column filled with NA
foo$d <- NA
replace elements with appropriate indexes with foo$a
foo$d[ind] <- foo$a[ind]

library(dplyr)
library(purrr)
# example dataset
foo<-data.frame(t = 1:15,
a = sample(1:15),
b = c(1,1,1,1,1,4,4,4,4,1,1,1,1,1,1),
c = sample(1:15))
# function to get indices of interest
# for a given index x go 2 positions back and 3 forward
# keep only positive indices
GetIDsBeforeAfter = function(x) {
v = (x-2) : (x+3)
v[v > 0]
}
foo %>% # from your dataset
filter(b > 1) %>% # keep rows where b > 1
pull(t) %>% # get the positions
map(GetIDsBeforeAfter) %>% # for each position apply the function
unlist() %>% # unlist all sets indices
unique() -> ids_to_remain # keep unique ones and save them in a vector
foo$d = foo$c # copy column c as d
foo$d[-ids_to_remain] = NA # put NA to all positions not in our vector
foo
# t a b c d
# 1 1 5 1 8 NA
# 2 2 6 1 14 NA
# 3 3 4 1 10 NA
# 4 4 1 1 7 7
# 5 5 10 1 5 5
# 6 6 8 4 9 9
# 7 7 9 4 15 15
# 8 8 3 4 6 6
# 9 9 7 4 2 2
# 10 10 12 1 3 3
# 11 11 11 1 1 1
# 12 12 15 1 4 4
# 13 13 14 1 11 NA
# 14 14 13 1 13 NA
# 15 15 2 1 12 NA

How to refer to multiple previous rows in R data.table

I have a question regarding data.table in R
i have a dataset like this
data <- data.table(a=c(1:7,12,32,13),b=c(1,5,6,7,8,3,2,5,1,4))
a b
1: 1 1
2: 2 5
3: 3 6
4: 4 7
5: 5 8
6: 6 3
7: 7 2
8: 12 5
9: 32 1
10: 13 4
Now i want to generate a third column c, which gonna compare the value of each row of a, to all previous values of b and check if there is any value of b is bigger than a. For e.g, at row 5, a=5, and previous value of b is 1,5,6,7. so 6 and 7 is bigger than 5, therefore value of c should be 1, otherwise it would be 0.
The result should be like this
a b c
1: 1 1 NA
2: 2 5 0
3: 3 6 1
4: 4 7 1
5: 5 8 1
6: 6 3 1
7: 7 2 1
8: 12 5 0
9: 32 1 0
10: 13 4 0
I tried with a for loop but it takes a very long time. I also tried shift but i can not refer to multiple previous rows with shift. Anyone has any recommendation?

library(data.table)
data <- data.table(a=c(1:7,12,32,13),b=c(1,5,6,7,8,3,2,5,1,4))
data[,c:= a <= shift(cummax(b))]

This is a base R solution (see the dplyr solution below):
data$c = NA
data$c[2:nrow(data)] <- sapply(2:nrow(data), function(x) { data$c[x] <- any(data$a[x] < data$b[1:(x-1)]) } )
## a b c
## 1: 1 1 NA
## 2: 2 5 0
## 3: 3 6 1
## 4: 4 7 1
## 5: 5 8 1
## 6: 6 3 1
## 7: 7 2 1
## 8: 12 5 0
## 9: 32 1 0
## 10: 13 4 0
EDIT
Here is a simpler solution using dplyr
library(dplyr)
### Given the cumulative max and comparing to 'a', set see to 1/0.
data %>% mutate(c = ifelse(a < lag(cummax(b)), 1, 0))
## a b c
## 1 1 1 NA
## 2 2 5 0
## 3 3 6 1
## 4 4 7 1
## 5 5 8 1
## 6 6 3 1
## 7 7 2 1
## 8 12 5 0
## 9 32 1 0
## 10 13 4 0
### Using 'shift' with dplyr
data %>% mutate(c = ifelse(a <= shift(cummax(b)), 1, 0))

calculate each chunk by group using dplyr?

How can I get the expected calculation using dplyr package?
row value group expected
1 2 1 =NA
2 4 1 =4-2
3 5 1 =5-4
4 6 2 =NA
5 11 2 =11-6
6 12 1 =NA
7 15 1 =15-12
I tried
df=read.table(header=1, text=' row value group
1 2 1
2 4 1
3 5 1
4 6 2
5 11 2
6 12 1
7 15 1')
df %>% group_by(group) %>% mutate(expected=value-lag(value))
How can I calculate for each chunk (row 1-3, 4-5, 6-7) although row 1-3 and 6-7 are labelled as the same group number?

Here is a similar approach. I created a new group variable using cumsum. Whenever the difference between two numbers in group is not 0, R assigns a new group number. If you have more data, this approach may be helpful.
library(dplyr)
mutate(df, foo = cumsum(c(T, diff(group) != 0))) %>%
group_by(foo) %>%
mutate(out = value - lag(value))
# row value group foo out
#1 1 2 1 1 NA
#2 2 4 1 1 2
#3 3 5 1 1 1
#4 4 6 2 2 NA
#5 5 11 2 2 5
#6 6 12 1 3 NA
#7 7 15 1 3 3

As your group variable is not useful for this, create a new variable aux and use it as the grouping variable:
library(dplyr)
df$aux <- rep(seq_along(rle(df$group)$values), times = rle(df$group)$lengths)
df %>% group_by(aux) %>% mutate(expected = value - lag(value))
Source: local data frame [7 x 5]
Groups: aux
row value group aux expected
1 1 2 1 1 NA
2 2 4 1 1 2
3 3 5 1 1 1
4 4 6 2 2 NA
5 5 11 2 2 5
6 6 12 1 3 NA
7 7 15 1 3 3

Here is an option using data.table_1.9.5. The devel version introduced new functions rleid and shift (default type is "lag" and fill is "NA") that can be useful for this.
library(data.table)
setDT(df)[, expected:=value-shift(value) ,by = rleid(group)][]
# row value group expected
#1: 1 2 1 NA
#2: 2 4 1 2
#3: 3 5 1 1
#4: 4 6 2 NA
#5: 5 11 2 5
#6: 6 12 1 NA
#7: 7 15 1 3

Resources