awk convert column matrix to square matrix?

johnpaulodonnell · 04-29-2008, 03:55 PM

Hi.

I have a 1*(N*N) column matrix and I need to convert it to a N*N matrix - each successive N entries of the column matrix becoming a column of the square matrix.

the sqrt of $NR number of records gives the dimension of the square matrix. Can I somehow change the record separator such that I can extract each successive N entries of the column matrix?

I can then format the matrix myself - the extraction of the columns is my problem!

Any help appreciated!

jschiwal · 04-29-2008, 05:19 PM

If you change the NF variable, you would probably simply drop the remaining fields.
I think different versions of awk behave differently when you change the NR variable. Especially if you increase it.

You could loop a variable from 1 to N*N and perform a modulas test ( NF % index ). If the result is zero, print \n.

Code:

echo >testfile {1..49}
~> echo >>testfile {1..36}
~> echo >>testfile {1..100}
~> awk 'BEGIN {ORS=" "}{ for (i=0; i<= NF; i++) {
                           if ((i%sqrt(NF) )==0) printf "\n"
                           print $(i+1)
                                                }}' testfile

1 2 3 4 5 6 7
8 9 10 11 12 13 14
15 16 17 18 19 20 21
22 23 24 25 26 27 28
29 30 31 32 33 34 35
36 37 38 39 40 41 42
43 44 45 46 47 48 49

1 2 3 4 5 6
7 8 9 10 11 12
13 14 15 16 17 18
19 20 21 22 23 24
25 26 27 28 29 30
31 32 33 34 35 36

1 2 3 4 5 6 7 8 9 10
11 12 13 14 15 16 17 18 19 20
21 22 23 24 25 26 27 28 29 30
31 32 33 34 35 36 37 38 39 40
41 42 43 44 45 46 47 48 49 50
51 52 53 54 55 56 57 58 59 60
61 62 63 64 65 66 67 68 69 70
71 72 73 74 75 76 77 78 79 80
81 82 83 84 85 86 87 88 89 90
91 92 93 94 95 96 97 98 99 100

I would have moved the sqrt() before the loop, but I was working interactively, which makes editing harder.

Code:

awk 'BEGIN {ORS=" "}{ dim=sqrt(NF)
                      for (i=0; i<= NF; i++) {
                           if ((i%dim )==0) printf "\n"
                           print $(i+1)
                                             }}' testfile

A test may be needed in case NF == 0, to avoid a division by zero error.

jschiwal · 04-29-2008, 06:33 PM

I may have misread your original problem. If each record is in it's own line like:
1
2
3
...
49

Then you can change the RS & FS variable in the BEGIN block so that the record is read in.

Code:

 cat testfile2
1
2
3
.
.
.
49

 awk 'BEGIN {RS="";FS="\n";ORS=" "}{ dim=sqrt(NF); for (i=0; i<= NF; i++) {
                           if ((i%dim )==0) printf "\n"
                           print $(i+1)
                                                }}' testfile2

1 2 3 4 5 6 7
8 9 10 11 12 13 14
15 16 17 18 19 20 21
22 23 24 25 26 27 28
29 30 31 32 33 34 35
36 37 38 39 40 41 42
43 44 45 46 47 48 49

johnpaulodonnell · 04-30-2008, 12:16 PM

Thanks - really appreciate your help!

NikosNL · 04-30-2008, 01:45 PM

hey thanks

i think i need this for this: http://www.linuxquestions.org/questi...e-bash-638748/

or not?

Thanks. When I am home i will try!