struct.unpack

[g.franzkowiak]

Hello Everybody,

I've read a pipe and store it in a object.
My next step was the separation from 4 bytes with
obj = string.join(lis t(dataObject)[:4] ==> '\x16 \x00 \x00 \x00'
and the converting by
value = struct.unpack(' I', obj) generated the error
"unpack str size does not match format"

Unfortunately is len(obj) 7, but integer lengt 4.
Why 7 ?

Any ideas ?

gf

[Fredrik Lundh]

Re: struct.unpack

"g.franzkow iak" wrote:
[color=blue]
> I've read a pipe and store it in a object.
> My next step was the separation from 4 bytes with
> obj = string.join(lis t(dataObject)[:4] ==> '\x16 \x00 \x00 \x00'
> and the converting by
> value = struct.unpack(' I', obj) generated the error
> "unpack str size does not match format"
>
> Unfortunately is len(obj) 7, but integer lengt 4.
> Why 7 ?[/color]

because string.join inserts a space between the bytes, by default (as
your example shows)

if you really need that join(list) thing (it's not clear how you read it, but
I find it a bit hard to believe that you've managed to read it into some-
thing that's supported by list but not unpack), you can do

obj = string.join(lis t(dataObject)[:4], "")

or

obj = "".join(list(da taObject[:4]))

or some variation thereof.

</F>

[g.franzkowiak]

Re: struct.unpack

Fredrik Lundh schrieb:[color=blue]
> "g.franzkow iak" wrote:
>
>[color=green]
>>I've read a pipe and store it in a object.
>>My next step was the separation from 4 bytes with
>>obj = string.join(lis t(dataObject)[:4] ==> '\x16 \x00 \x00 \x00'
>>and the converting by
>>value = struct.unpack(' I', obj) generated the error
>>"unpack str size does not match format"
>>
>>Unfortunate ly is len(obj) 7, but integer lengt 4.
>>Why 7 ?[/color]
>
>
> because string.join inserts a space between the bytes, by default (as
> your example shows)
>
> if you really need that join(list) thing (it's not clear how you read it, but
> I find it a bit hard to believe that you've managed to read it into some-
> thing that's supported by list but not unpack), you can do
>
> obj = string.join(lis t(dataObject)[:4], "")
>
> or
>
> obj = "".join(list(da taObject[:4]))
>
> or some variation thereof.
>
> </F>
>
>
>[/color]

I've also found the problem with som tests.
The best of this was:
tmpList = list(dataObject )[:4])
obj = tmpList[0]+tmpList[1]+tmpList[2]+tmpList[3].

But your suggestions are much better and new for me - both.

Thanks
gerd

[Peter Otten]

Re: struct.unpack

g.franzkowiak wrote:
[color=blue]
> tmpList = list(dataObject )[:4])
> obj = tmpList[0]+tmpList[1]+tmpList[2]+tmpList[3].[/color]

Have you tried just

obj = dataObject[:4]

without the intermediate list? If that failed, can you tell us the type of
the dataObject? E. g.
[color=blue][color=green][color=darkred]
>>> print type(dataObject )[/color][/color][/color]
<class '__main__.Neith erListNorString '>

Peter

[g.franzkowiak]

Re: struct.unpack

Peter Otten schrieb:[color=blue]
> g.franzkowiak wrote:
>
>[color=green]
>>tmpList = list(dataObject )[:4])
>>obj = tmpList[0]+tmpList[1]+tmpList[2]+tmpList[3].[/color]
>
>
> Have you tried just
>
> obj = dataObject[:4]
>
> without the intermediate list? If that failed, can you tell us the type of
> the dataObject? E. g.
>
>[color=green][color=darkred]
>>>>print type(dataObject )[/color][/color]
>
> <class '__main__.Neith erListNorString '>
>
> Peter
>[/color]

The dataObject was read from a named pipe as an byte stream

state, dataObject = win32file.ReadF ile(handle, nbytes, None)
print repr(dataObject )
==> '\x01\x02\x03\x 04\x00\x00\x00\ x00\x00\x00\x0. ....

With Frederiks help operates this fine

gerd

[Peter Otten]

Re: struct.unpack

g.franzkowiak wrote:
[color=blue]
> The dataObject was read from a named pipe as an byte stream
>
> state, dataObject = win32file.ReadF ile(handle, nbytes, None)
> print repr(dataObject )
> ==> '\x01\x02\x03\x 04\x00\x00\x00\ x00\x00\x00\x0. ....
>
> With Frederiks help operates this fine[/color]

I do not doubt that. My point was that instead of the suggested

# Fredrik
obj = "".join(list(da taObject[:4]))

the simpler

# me
obj = dataObject[:4]

might work as well. I cannot test it here, but judging from

http://aspn.activestate.com/ASPN/docs/ActivePython/2.3/pywin32/win32file__ReadFile_meth.html

the resulting dataObject is a buffer and buffer slices seem to be just
strings.
[color=blue]
> print repr(dataObject )
> ==> '\x01\x02\x03\x 04\x00\x00\x00\ x00\x00\x00\x0. ....[/color]

Hmm, that looks as if dataObject were a string -- please post the result of

print type(dataObject )

just to help me restore my peace of mind :-)

Peter

[g.franzkowiak]

Re: struct.unpack

Peter Otten schrieb:[color=blue]
> g.franzkowiak wrote:
>
>[color=green]
>>The dataObject was read from a named pipe as an byte stream
>>
>>state, dataObject = win32file.ReadF ile(handle, nbytes, None)
>>print repr(dataObject )
>> ==> '\x01\x02\x03\x 04\x00\x00\x00\ x00\x00\x00\x0. ....
>>
>>With Frederiks help operates this fine[/color]
>
>
> I do not doubt that. My point was that instead of the suggested
>
> # Fredrik
> obj = "".join(list(da taObject[:4]))
>
> the simpler
>
> # me
> obj = dataObject[:4]
>
> might work as well. I cannot test it here, but judging from
>
> http://aspn.activestate.com/ASPN/doc...File_meth.html
>
> the resulting dataObject is a buffer and buffer slices seem to be just
> strings.
>
>[color=green]
>>print repr(dataObject )
>> ==> '\x01\x02\x03\x 04\x00\x00\x00\ x00\x00\x00\x0. ....[/color]
>
>
> Hmm, that looks as if dataObject were a string -- please post the result of
>
> print type(dataObject )
>
> just to help me restore my peace of mind :-)
>
> Peter
>[/color]

Hello Peter,

was a node in my mind, the data comes as a string (with readfile always)
<< type(dataObject ) = 'str' >> and
<< type(dataObject :4) = 'str' >> also ;-)

I've dropped the loop with list/join and the result is the same.
Thank you :-))

gerd