python 2.3, cvs module specific question

Re: python 2.3, cvs module specific question

Bernard Delmée wrote:
[color=blue]
> Hello, I can't seem to be able to specify the delimiter when building a
> DictReader()
>[/color]

DictReader currently does not accept parameters for creating a dialect on
the fly. See Bug #792558.
[color=blue]
> One rather convoluted workaround is the following:
>
> inf = file('data.csv' )
> d = csv.Sniffer().s niff(s)
> inf.seek(0)
> headers = inf.readline(). split(';')
> rd = csv.DictReader( inf, headers, dialect=d )
> for row in rd:
> # ...
>[/color]

Another workaround might be (untested)

class MyDialect(csv.e xcel):
delimiter = ";"

inf = file('data.csv' )
headers = inf.readline(). split(';')
rd = csv.DictReader( inf, headers, dialect=MyDiale ct)

Of course your version should be more general, especially if you replace the
literal ';':

headers = inf.readline(). split(d.delimit er)


Peter

Re: python 2.3, cvs module specific question

Bernard> But this is rejected:

Bernard> inf = file('data.csv' )
Bernard> headers = inf.readline(). split(';')
Bernard> rd = csv.DictReader( inf, headers, delimiter=';' )
Bernard> for row in rd:
Bernard> # ...

This is a known problem. For the time being, define a new dialect class and
pass it in. If you know what the delimiter is, it's easier to define a
class than call the sniffer.

class semi(csv.excel) :
delimiter = ';'

inf = file('data.csv' )
headers = inf.readline(). split(';')
rd = csv.DictReader( f, headers, dialect=semi)
for row in rd:
...

Bernard> If DialectReader does indeed not accept the optional "fmtparam"
Bernard> then at least the documentation needs fixing. But then again I
Bernard> may just be misreading it....

No, it's just a bug in my implementation.

Skip

Re: python 2.3, cvs module specific question

> class semi(csv.excel) :[color=blue]
> delimiter = ';'[/color]

Thanks, guys, I hadn't thought of this simple solution.
It solves my problem nicely.
[color=blue]
> No, it's just a bug in my implementation.[/color]

This module as it is, is very useful, Skip. Frankly with
the speed improvements, the csv, datetime and logging
modules, python 2.3 is a really satisfying release.
Rest assured that your effort does not go unnoticed.

Cheers,

Bernard.