transposing square matrixs or blocks in a big file

Shell Programming and Scripting
1

Hi

I do have a big file of the following format

  a b c d e f g
  2 3 5 6 6 6 7
  3 4 5 6 7 9 0
  4 5 7 8 9 9 0
  1 2 4 5 6 7 8
  3 5 6 7 2 3 4
  5 6 7 4 3 2 4
  5 4 5 6 3 5 5 
  r h i j k l m 
  2 3 4 5 6 7 8
  4 5 7 8 9 9 0
  3 5 6 7 2 3 4
  2 3 5 6 6 6 7
  5 5 7 8 9 2 3
  1 2 4 5 6 7 8
  3 5 6 7 2 3 4

Basically the data is in a matrix format with many square matrixes put together. The bold fonts represents the beginning of such square matrixes. I have shown only 2 such blocks here but I have 100's of such blocks in that file in tab delimited format.

I want to to a transpose (change the axes) of each of the this square matrixes and get a tab delimted file as shown below.

a 2 3 4 1 3 5 5
b 3 4 5 2 5 6 4
c 5 5 7 4 6 7 5
d 6 6 8 5 7 4 6
e 6 7 9 6 2 3 3
f 6 9 9 7 3 2 5
g 7 0 0 8 4 4 5
r 2 4 3 2 5 1 3
h 3 5 5 3 5 2 5
i 4 7 6 5 7 4 6
j 5 8 7 6 8 5 7
k 6 9 2 6 9 6 2
l 7 9 3 6 2 7 3
m 8 0 4 7 3 8 4

Please let me know the best way to do it eith in sed or awk.

LA

---------- Post updated at 05:03 PM ---------- Previous update was at 03:39 PM ----------

Please let me know if there is a way to do this

2

Do the headers only contain letters and the fields only numbers?

3

The letters contains both letters and numbers (like letters_numbers) but the fields are always numbers.

Hope this helps.
LA

---------- Post updated at 05:17 PM ---------- Previous update was at 05:16 PM ----------

Sorry a typo

The heads (not letters) contains both letters and numbers (like letters_numbers) but the fields are always numbers.

4

You can just transpose the small blocks and print and then get the new block, not very memory intensive. What have you tried so far?

5

try:

 
 awk '
{if(/[a-z]/){x++}
{for(i=1;i<=NF;i++) a[x,NR,i]=$i;t=NR;s=NF
}}
END {for(l=1;l<=x;l++)
{for(m=1;m<=s;m++)
{for(n=1;n<=t;n++) {printf (length(a[l,n,m])>0?a[l,n,m]" ":"")}
{print ""}}}
}' urfile
1 Like
6

I can transpose block by block but there is like 800 blocks stagged in the file. I don't know how to loop it through.

---------- Post updated at 05:42 PM ---------- Previous update was at 05:32 PM ----------

Hi yinyuemi,

I tested with ur suggestion but its doing it only for the first 2 columns of each block (square matrix). If we could extend it to the 5 more columns in each block that would solve the problem.

7

Try this, which works as described earlier:

awk 'function pm(){
       for(j=1;j<=NF;j++)
         for(i=1;i<=m;i++)
           printf "%s"(i<m?OFS:RS),M[j,i]
     }
     /[[:alpha:]]/{
       if(m)pm()
       m=0
     }
     {
       ++m
       for(i=1;i<=NF;i++)
         M[i,m]=$i
     }
     END{
       pm()
     }' infile
1 Like
8

Thanks. I got it working. There was some issue in my data and after correcting it the first awk command itself worked.