UNIX for Dummies Questions & Answers

I want to find the different column numbers among rows in a file. For example:

A001 a b c d e ... N
A002 a b c d e ... N
A003 a b c d e ... N+1
A004 a b c d e ... N
A005 a b c d e ... N+2
:
:

For most of the lines I will have N columns (say 1000 rows) in each line except the line 3 and 5 (say 50 rows). I want to find all rows in my data file that contains different numbers of column, and print it out as

A003 N+1
A005 N+2

also, I want to delete those rows with different column numbers, what should I do for the command?

I use head awk '{print NF}' filename to find out the column numbers for each row, but there're just too many and always print on the screen at the same time~

How do you define "different" numbers of columns....
Anyway this reports the distribution of columns using space as the column delimiter:

Code:

awk '{ arr[NF]++}
       END { for (i in arr) {print i, arr[i]} } ' inputfilename

to remove offending rows you have to know what the column count should be:

Code:

echo "allowed number of columns \c"
read ccnt
awk -v ccnt=$ccnt ' NF==ccnt' inputfilename > newfilename

[/code]

To define the "different" numbers of columns in my dataset:

I know most of my rows in the dataset suppose to be N, but there are fews with N+2 or N+4, I want to find out which rows they are and delete it~

cat file | egrep "N\+[1-9]"

Code:

$ cat data.txt
A0001 1 2 3 4 5
A0002 1 2 3 4 5
A0003 1 2 3 4 5 6
A0004 1 2 3 4 5
A0005 1 2 3 4 5
A0006 1 2 3 4 5 6 7
A0007 1 2 3 4 5
A0008 1 2 3 4 5 6 7 8 9
A0009 1 2 3 4 5
A0010 1 2 3 4
$ awk '{ if(NF != 6){ print $1, NF } else { print >> "data2.txt" } }' data.txt
A0003 7
A0006 8
A0008 10
A0010 5
$ cat data2.txt
A0001 1 2 3 4 5
A0002 1 2 3 4 5
A0004 1 2 3 4 5
A0005 1 2 3 4 5
A0007 1 2 3 4 5
A0009 1 2 3 4 5
$

Thank you everyone! The codes are great and works!