Subsetting for matrices - r

Question

I'm playing around with some data, my matrix looks as shown in results:
results
[,1] [,2] [,3]
[1,] 1.7 0.0015 -1566.276
[2,] 1.7 0.0016 -1564.695
[3,] 1.7 0.0017 -1564.445
[4,] 1.7 0.0018 -1565.373
[5,] 1.7 0.0019 -1567.352
[6,] 1.7 0.0020 -1570.274
[7,] 1.8 0.0015 -1568.299
[8,] 1.8 0.0016 -1565.428
[9,] 1.8 0.0017 -1563.965
[10,] 1.8 0.0018 -1563.750
[11,] 1.8 0.0019 -1564.647
[12,] 1.8 0.0020 -1566.544
[13,] 1.9 0.0015 -1571.798
[14,] 1.9 0.0016 -1567.635
[15,] 1.9 0.0017 -1564.960
[16,] 1.9 0.0018 -1563.602
[17,] 1.9 0.0019 -1563.418
[18,] 1.9 0.0020 -1564.289
[19,] 2.0 0.0015 -1576.673
[20,] 2.0 0.0016 -1571.220
[21,] 2.0 0.0017 -1567.332
[22,] 2.0 0.0018 -1564.831
[23,] 2.0 0.0019 -1563.566
[24,] 2.0 0.0020 -1563.410
I wanted to print the row where the maximum on the 3rd column occurs, so I split it down in the following 2 lines:
max(results[,3])
results[results[,3] == -1563.41,]
The result didn't give me the desired row.
I tried putting the value in inverted commas as so:
max(results[,3])
results[results[,3] == "-1563.41",]
but this didn't work either.
the only code which worked was when I nested the max line of code in the 2nd line of code as so:
results[results[,3] == max(results[,3]),]
could someone please explain why breaking it down into steps didn't work??
I tried turning it into a data frame which didn't work either. using the filter from the tidyverse package didn't work either.
thanks,

Answer 1

R prints numbers to a maximum of getOption("digits") significant digits. I suspect that -1563.410 is the maximum rounded to seven significant digits, not the exact maximum. If no element of results[, 3] is exactly equal to the rounded number -1563.410, then
results[results[, 3] == -1563.41, ]
returns a 0-by-3 matrix. Meanwhile, max(results[, 3]) is the exact maximum, so
results[results[, 3] == max(results[, 3]), ]
returns the matrix subset that you want.
Note that if there are n rows whose third element is equal to max(results[, 3]), then the subset is a length-3 vector if n = 1 and an n-by-3 matrix if n > 1 (see ?Extract). If, in the n > 1 case, you only want the first of the n rows, then you should do:
results[which.max(results[, 3]), ]