eval-csv.py

#!/usr/bin/env python

from csv import DictReader, DictWriter
import sys
import re
from pprint import pprint
from datetime import datetime

parse_datetime = lambda x: datetime.strptime(x, '%Y-%m-%d %H:%M:%S')

types = { re.compile(r'^\d+$') : int, 
          re.compile(r'^[+-]?\ *(\d+(\.\d*)?|\.\d+)$') : float, 
          re.compile(r'^\d{4}-\d{2}-\d{2} \d{2}:\d{2}:\d{2}$') :  parse_datetime }
          
def parse_field(v):
    for pattern in types:
        if pattern.match(v):
            return types[pattern](v)

    return v

def main(option, args):

    f_in  = sys.stdin if args[0] == '-' else open(args[0], 'r')
    f_out = sys.stdout if args[1] == '-' else open(args[1], 'w')

    filter_func = (lambda x: eval(args[2], {}, x)) if len(args) > 2 else (lambda x: True)
    map_func    = (lambda x: eval(args[3], {}, x)) if len(args) > 3 else (lambda x: x)

    reader = DictReader(f_in)
    writer = DictWriter(f_out, reader.fieldnames)
         
    rows = ( dict( (k, parse_field(row[k])) for k in row ) for row in reader )

    writer.writerow( dict((i,i) for i in reader.fieldnames) )
    writer.writerows( map_func(row) for row in rows if filter_func(row) )

    return 0

if __name__ == '__main__':
    rv = main(None, sys.argv[1:])
    sys.exit(rv)