This is my dataset.
DTmtcars = as.data.table(mtcars); DTmtcars[, mtcarsname := rownames(mtcars)];
And these are the user input and can vary,
at least 1 will be always available which can be in any order,
these user inputs are just telling the column names of the data.table DTmtcars.
It is a sample for this question.
column1="mtcarsname";
column2="am";
column3="gear";
column4=NA;
requirement is display these columns of the dataset.
The code is an ideal solution for the sample user input above.
DTmtcars[, .(w,x,y) , env = list(w=column1, x = column2,y =column3,z = column4)]
probably an ifelse/fifelse or any thing else to make this code dynamic so that the output is produced for all possible combinations of user input.
Output should be data.table and not a vector or concat. code should be efficient and can be applied on a very large dataset. I would appreciate if it can be done using data.table package only.
As it seems you do not want the NA column out here is my try
library(data.table)
DTmtcars = as.data.table(mtcars); DTmtcars[, mtcarsname := rownames(mtcars)];
column1="mtcarsname";
column2="am";
column3="gear";
column4=NA;
DTmtcars[,.SD, .SDcols = na.omit(c(column1,column2,column3,column4))]
Which leads to
mtcarsname am gear
<char> <num> <num>
1: Mazda RX4 1 4
2: Mazda RX4 Wag 1 4
3: Datsun 710 1 4
4: Hornet 4 Drive 0 3
5: Hornet Sportabout 0 3
6: Valiant 0 3
7: Duster 360 0 3
8: Merc 240D 0 4
9: Merc 230 0 4
10: Merc 280 0 4
11: Merc 280C 0 4
....