如何在rpy2中创建一个两列矩阵

Question

I'm using rpy2 to run a method from an R library. According to the documentation:

method_name(x, range.x)

x: a two-column numeric matrix.
range.x: a list containing two vectors.

And it includes an example:

data(geyser, package="MASS")
x <- cbind(geyser$duration, geyser$waiting)
est <- method_name(x, range.x)

I checked the type of geyser$duration and geyser$waiting and both are double . I also tried replacing geyser$duration and geyser$waiting by g = c(.016, 2.15, 4.00) and h = c(.012, 2.11, 2.50) in R, and the code still works.

In my current Python code, I have:

import numpy as np
import rpy2.robjects as robjects
import rpy2.robjects.packages as rpackages
from rpy2.robjects.vectors import StrVector, FloatVector, ListVector # I tried these before too
from rpy2.robjects import numpy2ri, pandas2ri

numpy2ri.activate()
pandas2ri.activate()

base = rpackages.importr(('base'))

a = np.array([1.2, 2.1, 2.5]); b = np.array([5.2, 1.3, 2.15])
x = base.cbind(base.c(a), base.c(b))

ranges = base.range(x)
result = method_name(x, ranges)

As you can see, I'm trying to make my code as similar to the example as possible. However, I can't make the method work. I get the error Error in seq.default(a[2L], b[2L], length = M[2L]) which probably has to do with a problem in the arguments.

There's and obvious problem with ranges because it contains just two values, the minimum and maximum of x , however, it should contain two minimum values and to maximum values (one pair for each column of the matrix). I can achieve that by doing this:

ranges = base.cbind(base.range(a), base.range(b))

But this implies that there's a problem with the way I'm creating the matrix. Otherwise, I would get two pairs of values just by using base.range(x) .

I also tried x = robjects.r.matrix(x, ncol = 2) but didn't work. I still get just a global minimum and maximum value for the whole matrix when calling range .

What is the correct way of creating this matrix so that the method can run?

Answer 1

According to the documentation of the range function, it accepts as input vectors (one dimensional arrays). Thus, it would work by applying it to each column of your matrix or by applying it first directly to the a and b elements as you have mentioned. Thus, you second approach should work

# Define a,b vectors
a = np.array([1.2, 2.1, 2.5]); b = np.array([5.2, 1.3, 2.15])

# Calculate vector ranges
range_a = base.range(base.c(a))
range_b = base.range(base.c(b))

# Define the matrix 
x = base.cbind(base.c(a), base.c(b))
print(x)
>>> [[1.2  5.2 ]
    [2.1  1.3 ]
    [2.5  2.15]]

# Define the ranges
ranges = base.cbind(range_a, range_b)
print(ranges)
>>> [[1.2 1.3]
    [2.5 5.2]]