jpayne@11: #! /usr/bin/env python
jpayne@0: import csv
jpayne@0: import sys
jpayne@0: 
jpayne@0: 
jpayne@0: 
jpayne@0: def main(files):
jpayne@0: 	header = []
jpayne@0: 	items = []
jpayne@2: 	possible_identity_headers = None
jpayne@0: 	for fi in files:
jpayne@0: 		with open(fi, 'rU') as table:
jpayne@2: 			reader = csv.DictReader(table, delimiter='\t', dialect='excel-tab')
jpayne@2: 			rows = list(reader)
jpayne@2: 			for field in reader.fieldnames:
jpayne@0: 				if field not in set(header):
jpayne@0: 					header.append(field)
jpayne@2: 				#try to find identity columns in the files, to use to join
jpayne@2: 				if possible_identity_headers is None:
jpayne@2: 					possible_identity_headers = set(reader.fieldnames)
jpayne@2: 				#winnow down the shared columns in each file by whether they're present in all, and all their values are unique in each file and not null
jpayne@2: 				#because these are the most likely to be shared keys
jpayne@2: 				possible_identity_headers = possible_identity_headers.intersection(filter(lambda f: len(set([r[f] for r in rows])) == len(rows) and all([r[f] is not None for r in rows]), reader.fieldnames))
jpayne@0: 			items.extend(rows)
jpayne@2: 	
jpayne@2: 	# if len(possible_identity_headers) > 1:
jpayne@2: 	# 	#if there's more than one, we need to check that joining on any one of them produces the same results
jpayne@2: 
jpayne@2: 	# 	#finally
jpayne@2: 	# 	possible_identity_headers = set((possible_identity_headers.pop(), ))
jpayne@2: 
jpayne@2: 	#if we found an identity column, then try to join rows
jpayne@2: 	if possible_identity_headers:
jpayne@2: 		key_column = possible_identity_headers.pop()
jpayne@2: 		keys = set([r[key_column] for r in items])
jpayne@2: 		merged_rows = []
jpayne@2: 		for key in sorted(keys):
jpayne@2: 			new_row = {}
jpayne@2: 			for row in filter(lambda r: r[key_column] == key, items):
jpayne@2: 				new_row.update(row)
jpayne@2: 			merged_rows.append(new_row)
jpayne@2: 		items = merged_rows
jpayne@2: 
jpayne@0: 	wr = csv.DictWriter(sys.stdout, delimiter='\t', dialect='excel-tab', fieldnames=header)
jpayne@0: 	wr.writeheader()
jpayne@0: 	wr.writerows(items)
jpayne@0: 
jpayne@0: 
jpayne@0: if __name__ == '__main__':
jpayne@0: 	main(sys.argv[1:])