Turn vector output into columns in data.table along with other columns?

Question

My questions is related to this one:

Turn vector output into columns in data.table?

But my situation is a bit more complicated. I am not only returning the vector as the columns, but I am also calculating other columns at the same time. E.g.:

DT = data.table(X = 1:10, Y = 11:20, Z = 21:30, group = rep(1:10, each = 3))

featuresDT <- quote(list(x = mean(X),
                         y = mean(Y),
                         z = mean(Z),
                         as.list(quantile(X))))

DT[, eval(featuresDT), by = "group"]

where quantile returns a length 5 vector. Instead of getting a data.table with 8 columns, I am getting one with 4 columns and the quantile results are displayed as extra rows and x, y and z are duplicated 5 times. I also tried dist = as.list(quantile(X) but that gives the same result but different column name.

can you please provide a reproducible example ?

statquant
– statquant

2013-08-12 11:36:03 +00:00
Commented Aug 12, 2013 at 11:36 — statquant
– statquant, Commented Aug 12, 2013 at 11:36

Arun · Accepted Answer · 2013-08-12 23:40:54Z

3

You should just change the list to c. c with any value of type list will automatically result in a list):

featuresDT <- quote(c(x = mean(X),
                         y = mean(Y),
                         z = mean(Z),
                         as.list(quantile(X))))
DT[, eval(featuresDT), by = "group"]

    group        x        y        z 0% 25% 50% 75% 100%
 1:     1 2.000000 12.00000 22.00000  1 1.5   2 2.5    3
 2:     2 5.000000 15.00000 25.00000  4 4.5   5 5.5    6
 3:     3 8.000000 18.00000 28.00000  7 7.5   8 8.5    9
 4:     4 4.333333 14.33333 24.33333  1 1.5   2 6.0   10
 5:     5 4.000000 14.00000 24.00000  3 3.5   4 4.5    5
 6:     6 7.000000 17.00000 27.00000  6 6.5   7 7.5    8
 7:     7 6.666667 16.66667 26.66667  1 5.0   9 9.5   10
 8:     8 3.000000 13.00000 23.00000  2 2.5   3 3.5    4
 9:     9 6.000000 16.00000 26.00000  5 5.5   6 6.5    7
10:    10 9.000000 19.00000 29.00000  8 8.5   9 9.5   10

answered Aug 12, 2013 at 23:40

Arun

119k28 gold badges290 silver badges396 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

Farrel Over a year ago

quantile function gives each out a name (eg "0%", "25%"). That's nice. What about arguments that output a vector but one that is unamed (eg range())?. Could I specify names?

Roland · Accepted Answer · 2013-08-12 12:01:49Z

1

Try this:

featuresDT <- quote(cbind(list(x = mean(X),
                         y = mean(Y),
                         z = mean(Z)),
                         as.data.table(t(quantile(X)))))

answered Aug 12, 2013 at 12:01

Roland

134k12 gold badges203 silver badges305 bronze badges

3 Comments

mchangun Over a year ago

That works - thank you! Is there a way for me to change the colnumn names? I am doing quantile for X Y and Z and the present solution gives me duplicated column names.

Frank Over a year ago

@mchangun Just write X = as.data.table(... and an "X." will prefix the current column names.

mchangun Over a year ago

@Roland I modified your code slightly - I am doing:

featuresDT <- quote(c(list(x = mean(X),                                y = mean(Y),                                z = mean(Z)),                           X = as.list(quantile(X))))

Is there any performance difference the two solutions? Keeping everything as lists seem to be more in line with the "spirit" of data.tables.

Collectives™ on Stack Overflow

Turn vector output into columns in data.table along with other columns?

2 Answers 2

1 Comment

3 Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

1 Comment

3 Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related