CODEWITHSUNDEEP
rdata.tablesubset
bsg
bsg

Reputation: 835

Use variable in i of data.table subset

I am new to data.table and think that this is an easy question, but can't seem to find the answer anywhere.

I want to subset a table based on the value of two columns, whose names I know. But I want to compare against a value which I don't know in advance. That is, I want to use a variable for the i portion of DT[]. But I can't seem to figure out how to do it. Everything I see explains how to use a variable for j (i.e. column names), but not for i.

When I just put the name of the variable in, i.e.

setkey(dtpredictions, colA, colB)
nextweek = dtpredictions[J(uservar, weekvar)]

it returns the entire table. Trying to apply the answer to FAQ 1.6, I tried:

  nextweek = dtpredictions[J(eval(quote(uservar)), eval(quote(weekvar)))]

and 

 nextweek = dtpredictions[J(eval(user), eval(week))]

but both still returned the whole table.

I am pretty sure this is very simple, but I am stuck.

EDIT I apologize for not clarifying earlier: I would like to do a binary search, since I need the speedup. I know that I can do a vector scan using ==, but I would prefer not to.

Upvotes: 3

Views: 1139

Answers (3)

bsg
bsg

Reputation: 835

Found the problem - one of my variables had the same name as a column in the table. I actually saw a question about a similar problem here, but didn't even realize that I had that issue. (It was another column in the table, not the one I was subsetting on.)

I changed the name of the variable I was using to subset and now it works.

Upvotes: 3

Mike.Gahan
Mike.Gahan

Reputation: 4615

hmmm...interesting. Does this code seem to work for you? I am not getting the same error. I am using data.table 1.9.3.

require(data.table)
iris <- data.table(iris)

#Create new categorical variable
set.seed(1)
iris[ , new.var := sample(letters[1:5],150,replace=TRUE)]

#Set keys
setkey(iris,Species,new.var)

#Create variables to reference
check1 <- "setosa"
check2 <- "b"

#Return matches
iris[J(check1,check2)]

And the resulting table:

    Sepal.Length Sepal.Width Petal.Length Petal.Width Species new.var
 1:          5.1         3.5          1.4         0.2  setosa       b
 2:          4.9         3.0          1.4         0.2  setosa       b
 3:          5.0         3.6          1.4         0.2  setosa       b
 4:          5.4         3.7          1.5         0.2  setosa       b
 5:          4.3         3.0          1.1         0.1  setosa       b
 6:          5.7         3.8          1.7         0.3  setosa       b
 7:          5.1         3.7          1.5         0.4  setosa       b
 8:          4.8         3.4          1.9         0.2  setosa       b
 9:          5.0         3.0          1.6         0.2  setosa       b
10:          5.2         3.5          1.5         0.2  setosa       b
11:          4.7         3.2          1.6         0.2  setosa       b

Upvotes: 1

djhurio
djhurio

Reputation: 5536

Is this you are looking for?

setkey(dtpredictions, colA, colB)
nextweek <- dtpredictions[colA == uservar & colB == weekvar]

Upvotes: 0

Related Questions