Project

General

Profile

« Previous | Next » 

Revision 3822

Added by Aaron Marcuse-Kubitza over 12 years ago

join: Support discarding multiple outputs if they should be considered ambiguous

View differences:

join
1 1 
#!/usr/bin/env python
2 2 
# Inner-joins two map spreadsheets A->B and B->C to A->C
3 
# Multi-safe (supports an input appearing multiple times).
3 
# Multi-safe (supports an input appearing multiple times). If multiple outputs
4 
# should be considered ambiguous, they can be discarded by setting only_one.
4 5 

  		




  5
  6
  
    
import csv


  		




  6
  7
  
    
import operator


  		




  ......




  11
  12
  
    
sys.path.append(os.path.dirname(__file__)+"/../lib")


  		




  12
  13
  
    

  		




  13
  14
  
    
import maps


  		




  
  15
  
    
import opts


  		




  14
  16
  
    
import util


  		




  15
  17
  
    

  		




  16
  18
  
    
def main():


  		




  
  19
  
    
    only_one = opts.env_flag('only_one')


  		




  17
  20
  
    
    try: _prog_name, map_1_path = sys.argv


  		




  18
  21
  
    
    except ValueError:


  		




  19
  
  
    
        raise SystemExit('Usage: '+sys.argv[0]+' <map_0 map_1 [| '+sys.argv[0]


  		




  20
  
  
    
            +' map_2]... >joined_map')


  		




  
  22
  
    
        raise SystemExit('Usage: env [only_one=1] '+sys.argv[0]


  		




  
  23
  
    
            +' <map_0 map_1 [| '+sys.argv[0]+' map_2]... >joined_map')


  		




  21
  24
  
    
    


  		




  22
  25
  
    
    # Get map 1


  		




  23
  26
  
    
    map_1 = {}


  		




  ......




  64
  67
  
    
            


  		




  65
  68
  
    
            # Write new mapping


  		




  66
  69
  
    
            is_empty = len(out_rows) == 1 and out_rows[0][1] == ''


  		




  67
  
  
    
            if out_rows and not is_empty: # found non-empty mapping(s)


  		




  
  70
  
    
            if only_one and len(out_rows) > 1: # multiple outputs are ambiguous


  		




  
  71
  
    
                set_error('Ambiguous mapping for '+row[1]) # discards mapping


  		




  
  72
  
    
            elif out_rows and not is_empty: # found non-empty mapping(s)


  		




  68
  73
  
    
                for out_row in out_rows:


  		




  69
  74
  
    
                    row_ = row[:] # don't modify row, since it will be reused


  		




  70
  75
  
    
                    row_ = maps.merge_mappings(row_, out_row)


  		












Also available in:  Unified diff