Project

General

Profile

« Previous | Next » 

Revision 5145

Added by Aaron Marcuse-Kubitza over 12 years ago

csvs.py: TsvReader: Also interpret '\t' as a tab, to provide a mechanism for encoding embedded tabs

View differences:

lib/csvs.py
38 38 

  		




  39
  39
  
    
class TsvReader:


  		




  40
  40
  
    
    '''Unlike csv.reader, for TSVs, interprets \ as escaping a line ending but


  		




  41
  
  
    
    ignores it before everything else (e.g. \N for NULL). Also interprets the


  		




  42
  
  
    
    '\n' escape sequence as a newline.'''


  		




  
  41
  
    
    ignores it before everything else (e.g. \N for NULL).


  		




  
  42
  
    
    Also interprets '\n' as a newline  and '\t' as a tab.


  		




  
  43
  
    
    '''


  		




  43
  44
  
    
    def __init__(self, stream, dialect):


  		




  44
  45
  
    
        assert is_tsv(dialect)


  		




  45
  46
  
    
        self.stream = stream


  		




  ......




  64
  65
  
    
        record = record.replace('\r', ending_placeholder)


  		




  65
  66
  
    
        


  		




  66
  67
  
    
        row = csv.reader(StringIO.StringIO(record), self.dialect).next()


  		




  67
  
  
    
        return [v.replace(ending_placeholder, '\n') for v in row]


  		




  
  68
  
    
        return [v.replace(r'\n', '\n').replace(r'\t', '\t') for v in row]


  		




  68
  69
  
    

  		




  69
  70
  
    
def reader_class(dialect):


  		




  70
  71
  
    
    if is_tsv(dialect): return TsvReader


  		












Also available in:  Unified diff