Grdmath question - filling an (n x m) grid with a column vector or a row vector

Q&A
1

Strange problem that likely has an easy solution, but it eludes me.

I need to create a full (n x m) grid based on an (n x 1) column vector. In other words, the column vector is repeated for each m column.

Simple example, given a column vector:
| 1 |
| 2 |
| 3 |,
create a matrix of repeated columns:
| 1 1 1 1 |
| 2 2 2 2 |
| 3 3 3 3 |

I also need to create a full (n x m) grid based on a (1 x m) row vector. In other words, the row is repeated for each n row.

Example, given a row:
| 1 2 3 |,
create a matrix of repeated rows:
| 1 2 3 |
| 1 2 3 |
| 1 2 3 |
| 1 2 3 |

Is it possible to fool grdmath to do either of these in one command, given -R & -I parameters that correspond to the (m x n) grid?

In my case, I’m not working with sequential integers, but with mixed values read-in as a column or as a row.

One possible solution is to read-in the column (or row), then repeat it (n or m number of times, depending on case) into a temporary file (via cat >>), then use xyz2grd with the appropriate -Z specification to form the grid. Maybe that’s the easiest way - since the final grids aren’t that huge.

2

Replicating a column is simple. Here t.txt is your 1, 2, 3 column vector file:

 gmt math -N4 t.txt 0 COL ADD =
    1 1 1 1
    2 2 2 2
    3 3 3 3

To make a matrix I think you are better off just awking out as many as you need

awk ‘{for (i = 1; i <= 50; i++) print $0}’ row.txt

3

Appreciate your reply, Paul. So, gmt math can take any nx1 column vector, and form an nxm matrix, very quickly. That is useful!

And the awk command can easily repeat a row. Good stuff. Now I’m unable to locate the script where I was needing to do this! Too long ago! Ha!