ncount

Synopsis

A trivial command-line program to print lines from a delimited file that don't match a given field count.

ncount -h

Usage: ncount [OPTION]... [FILE]...
Output records from a delimited file NOT matching the given field count.
More than one FILE can be specified.

  -d, --delimiter=DELIM  the delimiting character for the input FILE(s)
  -n, --field-count=FC   the field count to use while processing (required)
  -l, --add-line         include the line number in the output
  -c, --add-count        include the field count in the output
  -C  --csv              parse CSV files
  -Q, --csv-quote        CSV quoting character (ignored unless --csv)
  -N, --csv-nl-count     output CSV records with embedded newlines
  -h, --help             This help

Building ncount

The following projects have made ncount possible:

• libcsv - Version 3.0.3 of libcsv is included with ncount.
• gnulib - The getline module is included with ncount for portability.

Please consider contributing to those projects if you find ncount useful.

Note that this git repository does not include a configure script like a distribution tarball normally does. If you don't want to bother with installing autoconf and automake, then download a distribution package and run configure:

./configure

If you want the ncount binary installed in your $HOME/bin, you should run something like:

./configure --prefix=$HOME

If you're building from a copy of this git repository, you'll need to have autoconf and automake installed on your system, and run the following command to generate a configure script:

autoreconf -i

...and subsequently run configure as mentioned before.

After configure completes successfully, you can do the usual:

make
make check
make install

Why?

Speed.

time awk -F$'\t' '{if ( NF != 19 ) print $0}' 50_million_row_file.txt

real    1m59.619s
user    1m58.566s
sys     0m1.052s

time ncount -n 19 50_million_row_file.txt

real    0m17.425s
user    0m12.554s
sys     0m4.871s

Author

Miguel Gualdron (dev at gualdron.com).