Shell Programming and Scripting

it didn't work. there are uneven values in eash row, some row have 500 records, some have 200, i dont' know if that's problem.

the error is,

bash-2.03$ ./file_transpose
nawk: record `atcc_rma_pvsall_bot5...' has too many fields
input record number 3, file atcc_fidler_genename_markers
source line number 2

the code is,

###file transpose, colume to row
###from unix.com/Ygor, 07/29/2005

nawk -v OFS='\t' '{
for (f = 1; f <= NF; f++)
a[NR, f] = $f
}
NF > nf { nf = NF }
END {
for (f = 1; f <= nf; f++)
for (r = 1; r <= NR; r++)
printf a[r, f] (r==NR ? RS : OFS)
}' infile > outfile

that is the problem - too many fields for the 'nawk' to handle.
if you have 'gawk' installed on your system - try that instead of 'nawk'.
or if on Solaris, try '/usr/xpg4/bin/awk'.

The other problem is that the solution assumes the consistent number of columns for ALL the rows.

Good luck.

P.S. for the future - pls don't repost already posted unmodified solutions - it clutters the thread.

it worked! thank you so much. but only one problem left, some of the record being changed. for example,

input record c2 is mad1///mad2///mad3 (one record), output is mad1, ///, mad2, ///, mad3 (file records); can this be fixed? thanks again.

I got your point for post a request. Thanks.

I don't quite understand what you're saying, but you'll have to look at your input data and see the 'wrong' records "differ" from the rest - most likely this "difference" breaks the implementation.

Good luck.