Balancing (creating the same number of lines for each individual user)

Given the data table as follows, id1

is a subject-level identifier, id2

is an identifier of a repeating measure within an object, X

are data variables, of which there are many. I want to balance the data in such a way that each person has the same number of rows (repeated measurements), which is max(DT[,.N,by=id1][,N])

, but where id1

and is id2

adjusted as needed, and X

the data values ​​are replaced with NA

for these new rows.

Following:

DT = data.table(
id1 = c(1,1,2,2,2,3,3,3,3),
id2 = c(1,2,1,2,3,1,2,3,4),
X1 = letters[1:9],
X2 = LETTERS[1:9]
)
setkey(DT,id1)

      

It should look like:

DT = data.table(
id1 = c(1,1,1,1,2,2,2,2,3,3,3,3),
id2 = c(1,2,3,4,1,2,3,4,1,2,3,4),
X1 = c(letters[1:2],NA,NA,letters[3:5],NA,letters[6:9]),
X2 = c(LETTERS[1:2],NA,NA,LETTERS[3:5],NA,LETTERS[6:9])
)

      

How do you do it using data.table

? To do this, you must avoid loops as this dataset is huge. Is this a job for reshape2

?

+3


source to share


2 answers


You may try:



 DT2 <- CJ(id1=1:3, id2=1:4)
 merge(DT,DT2, by=c('id1', 'id2'), all=TRUE)

      

+4


source


Here's a slight deviation from akrun's answer that is commonly used for this problem:



setkey(DT, id1, id2)
DT[CJ(unique(id1), unique(id2))]

      

+2


source







All Articles