comparison column_order_header_sort.py @ 0:6ae9724caf4d draft default tip

planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/column_order_header_sort commit d562cc65926c8c95af21467177b253b6ac985cb4
author iuc
date Wed, 12 Apr 2017 17:17:18 -0400
parents
children
comparison
equal deleted inserted replaced
-1:000000000000 0:6ae9724caf4d
1 #!/usr/bin/env python
2
3 import subprocess
4 import sys
5
6 AWK_CMD = """BEGIN{FS="%s"; OFS="%s";} {print %s;}"""
7
8 input_filename = sys.argv[1]
9 output_filename = sys.argv[2]
10 delimiter = sys.argv[3]
11 key_column = sys.argv[4]
12
13 try:
14 key_column = int( key_column ) - 1
15 except Exception:
16 key_column = None
17
18 header = None
19 with open( input_filename, 'r' ) as fh:
20 header = fh.readline().strip( '\r\n' )
21 header = header.split( delimiter )
22 assert len( header ) == len( set( header ) ), "Header values must be unique"
23 sorted_header = list( header )
24 if key_column is None:
25 columns = []
26 else:
27 columns = [ key_column ]
28 sorted_header.pop( key_column )
29 sorted_header.sort()
30
31 for key in sorted_header:
32 columns.append( header.index( key ) )
33
34 awk_cmd = AWK_CMD % ( delimiter, delimiter, ",".join( map( lambda x: "$%i" % ( x + 1 ), columns ) ) )
35 sys.exit( subprocess.call( [ 'gawk', awk_cmd, input_filename ], stdout=open( output_filename, 'wb+' ), shell=False ) )