Shell Programming and Scripting

sample filenames: 20000105.u, 20000112.u and so on
for files with the same numeric values in blue; do
average of col.3 
done

say all files have 3 columns by 5 rows: 
the mean i wanted would be: (col3row1 in file1 + col3row1 in file2)/2 to (col3row5 in file1 + col3row5 in file2)/2

ls *.u | cut -c1-6 | sort -u >list
while read a
do
nawk '{A[NR]=$1 FS $2;B[FNR]+=$3;++N[FNR]
}END{
i=1;do{ print A[i];}while(++i in A)
j=1;do{ print B[j]/N[j];}while(++j in B)
}' ${a}* >"$a.out"
done <list

(20000105.u,20000102.u,20000119.u,20000126.u) 

Each of these files contain say 3 cols and 3 rows for instance:
file with filename: 20000105.u contains
lon lat value 
123.0 12.0 0.5
123.1 12.0 1.4
123.2 12.0 1.5

file with filename: 20000112.u contains
lon lat value 
123.0 12.0 0.6
123.1 12.0 1.8
123.2 12.0 1.9

file with filename: 20000119.u contains
lon lat value 
123.0 12.0 0.8
123.1 12.0 1.8
123.2 12.0 2.0

file with filename: 20000126.u contains
lon lat value 
123.0 12.0 1.0
123.1 12.0 2.4
123.2 12.0 185

Expected output:
lon lat value 
 123.0 12.0 average of values from 4files for row1
 123.1 12.0 average of values from 4files for row2
 123.2 12.0 average of values from 4files for row3

awk '
NR==1{x=$0;next}
FNR>1{A[FNR]=$1 FS $2;B[FNR]+=$3;N[FNR]++;c=N[FNR]
}END{
print x;i=2;do {print A[i] "average of value from " c " files for row " i "=" B[i]/N[i] }while(++i in B)}' 200001*.u

ls *.u | cut -c1-6 | sort -u >list
while read a
do
awk 'NR==1{x=$0;next}
FNR>1{A[FNR]=$1 FS $2;B[FNR]+=$3;N[FNR]++;c=N[FNR]
}END{print x; i=2;do {print A[i] "average of value from " c " files for row " i "=" B[i]/N[i] }while(++i in B)}
' ${a}* >"$a.out"
done <list