I have to convert some txt files to csv (and make some operation during the conversion).

I use csv.Sniffer() class to detect wich delimiter is used in the txt

This code

with open(filename_input, 'r') as f1, open(filename_output, 'wb') as f2:dialect = csv.Sniffer().sniff(f1.read(1024)) #### detect delimiters  f1.seek(0)r=csv.reader(f1, delimiter=dialect )writer = csv.writer(f2,delimiter=';')

return: Error: Could not determine delimiter

This work

with open(filename_input, 'r') as f1, open(filename_output, 'wb') as f2:#dialect = csv.Sniffer().sniff(f1.read(1024)) #### detect delimiters  #f1.seek(0)r=csv.reader(f1, delimiter='\t' )writer = csv.writer(f2,delimiter=';')

or

with open(filename_input, 'r') as f1, open(filename_output, 'wb') as f2:#dialect = csv.Sniffer().sniff(f1.read(1024)) #### detect delimiters  #f1.seek(0)r=csv.reader(f1, dialect="excel-tab")writer = csv.writer(f2,delimiter=';')

this is a txt row example (10 records delimited by Tab)

166 14908941    sa_s    NOVA i  7.05    DEa 7.17    Ncava - Deo mo  7161    4,97

why csv.Sniffer() class doesn't work?

The bug was read only 1024 byte to parse the entire txt(maybe this is not enough to detect the delimiter). Now this code works without no other edit:

with open(filename_input, 'r') as f1, open(filename_output, 'wb') as f2:dialect = csv.Sniffer().sniff(f1.read()) #### error with dialect = csv.Sniffer().sniff(f1.read(1024))  f1.seek(0)r=csv.reader(f1, delimiter=dialect )writer = csv.writer(f2,delimiter=';')

You have to use dialect.delimiter instead of just dialect because what is returned is of type class Dialect and you need its attribute Dialect.delimiter

rows=csv.reader(f1, delimiter=dialect.delimiter)

Modified code will be as below

import csvfilename_input = 'filein.txt'
filename_output = 'fileout.csv'
with open(filename_input, 'r') as f1, open(filename_output, 'wb') as f2:dialect = csv.Sniffer().sniff(f1.read(1024), "\t") #### detect delimitersf1.seek(0)print(dialect.delimiter)rows=csv.reader(f1, delimiter=dialect.delimiter)writer = csv.writer(f2,delimiter=';')writer.writerows(rows)

Output:

C:\pyp>python.exe txttocsv.py
,
C:\pyp>

Also note that from doc:

sniff(sample, delimiters=None)

Analyze the given sample and return a Dialect subclass reflectingthe parameters found. If the optional delimiters parameter is given,it is interpreted as a string containing possible valid delimitercharacters.

Hence if the delimiter that you want to find in your text file is something like # instead of , or ; then you should mention that in sniff function as second parameter like this:

dialect = csv.Sniffer().sniff(f1.read(1024), '#') 

Update: For reading whole file you will need

dialect = csv.Sniffer().sniff(f1.read())