Revision 8202
Added by Aaron Marcuse-Kubitza over 11 years ago
| lib/csvs.py | ||
|---|---|---|
| 36 | 36 |
|
| 37 | 37 |
return dialect |
| 38 | 38 |
|
| 39 |
def has_unbalanced_quotes(str_): return str_.count('"') % 2 == 1 # odd # of "
|
|
| 40 |
|
|
| 41 |
def has_multiline_column(str_): return has_unbalanced_quotes(str_) |
|
| 42 |
|
|
| 39 | 43 |
def stream_info(stream, parse_header=False): |
| 40 | 44 |
'''Automatically detects the dialect based on the header line. |
| 41 | 45 |
Uses the Excel dialect if the CSV file is empty. |
| 42 | 46 |
@return NamedTuple {header_line, header, dialect}'''
|
| 43 | 47 |
info = util.NamedTuple() |
| 44 | 48 |
info.header_line = stream.readline() |
| 49 |
if has_multiline_column(info.header_line): # 1st line not full header |
|
| 50 |
# assume it's a header-only csv with multiline columns |
|
| 51 |
info.header_line += ''.join(stream.readlines()) # use entire file |
|
| 45 | 52 |
info.header = None |
| 46 | 53 |
if info.header_line != '': |
| 47 | 54 |
info.dialect = sniff(info.header_line) |
Also available in: Unified diff
lib/csvs.py: stream_info(): Fixed bug where headers with multiline columns were not supported because only the first line (not the first multiline row) is sniffed for the dialect