Prev: Creating a logical mask from a 3D convex hull
Next: About behavior of KeyPressFcn
From: dpb on 7 Jul 2010 16:00
Brought to light in another thread of poster w/ fixed width fields but
no delimiter wherein import wizard, etc, failed to parse the empty fields.

Thought should be able to use the width but no; at least w/ my early
version and as I read doc on current textscan() probably won't work
there, either.

So, is the following

a) broke in early release but fixed now, or

b) wad?

>> s = ' 3 12.5';
>> [x,y,z]=strread(s,'%3d%3d%3.1f',1)
x =
3
y =
12
z =
0.5000
>>

OTOH,

>> [x,y,z]=strread(strrep(s,' ','0'),'%3d%3d%3.1f')
x =
3
y =
1
z =
2.5000

as expected/desired.

So, is there any way to parse the "empty" fields that aren't really
empty w/o resorting to somesuch workaround as the above?

And, b) what is actually going on w/ the prior result--it makes no sense
to me to get y=12 z=0.5, certainly.

--
From: Oleg Komarov on 7 Jul 2010 16:24
dpb <none(a)non.net> wrote in message <i12mii$vth$1(a)news.eternal-september.org>...
> Brought to light in another thread of poster w/ fixed width fields but
> no delimiter wherein import wizard, etc, failed to parse the empty fields.
>
> Thought should be able to use the width but no; at least w/ my early
> version and as I read doc on current textscan() probably won't work
> there, either.
>
> So, is the following
>
> a) broke in early release but fixed now, or
>
> b) wad?
>
> >> s = ' 3 12.5';
> >> [x,y,z]=strread(s,'%3d%3d%3.1f',1)
> x =
> 3
> y =
> 12
> z =
> 0.5000
> >>
>
> OTOH,
>
> >> [x,y,z]=strread(strrep(s,' ','0'),'%3d%3d%3.1f')
> x =
> 3
> y =
> 1
> z =
> 2.5000
>
> as expected/desired.
>
> So, is there any way to parse the "empty" fields that aren't really
> empty w/o resorting to somesuch workaround as the above?
>
> And, b) what is actually going on w/ the prior result--it makes no sense
> to me to get y=12 z=0.5, certainly.
>
> --

I found this problem several times and apparently there is no way to parse empty fields (also confirmed by tech support) !!! I thought it was just my problem but since you bring it to our attention I can confirm that my solution was similar, resorting to replace and textscanning after.

Oleg
From: dpb on 7 Jul 2010 16:49
Oleg Komarov wrote:
....

> I found this problem several times and apparently there is no way to
> parse empty fields (also confirmed by tech support) !!! I thought it was
> just my problem but since you bring it to our attention I can confirm
> that my solution was similar, resorting to replace and textscanning after.
....

Oleg, out of curiosity do you happen to know if the behavior mimics C
scanf() formatting string behavior or is it unique to Matlab? I don't
know C well enough to have a clue (and am not sure there's a
working-at-the-moment C compiler handy to give a short test)...

--
From: Oleg Komarov on 7 Jul 2010 16:56
> I found this problem several times and apparently there is no way to parse empty fields (also confirmed by tech support) !!! I thought it was just my problem but since you bring it to our attention I can confirm that my solution was similar, resorting to replace and textscanning after.
>
> Oleg

I had a test.txt file:
0000000001901011943 IBO009900002002010142700EUR000000000166322000000000000000
0000000001901011943USABO009900002002010142700EUR000000000166422000000000000000

And the solution proposed by the support (commented in italian):
%% Apertura del file
fid = fopen('leggi.txt','r');

%% Lettura stringhe di riga
A = textscan(fid,'%s','delimiter','\n');
A = cell2mat(A{1});

%% Salvataggio stringhe
field1 = A(:,1:19);
field2 = A(:,20:22);

%% Chiusura del file
fclose(fid)

Oleg
From: Oleg Komarov on 7 Jul 2010 16:58
dpb <none(a)non.net> wrote in message <i12pe1$nkv$1(a)news.eternal-september.org>...
> Oleg Komarov wrote:
> ...
>
> > I found this problem several times and apparently there is no way to
> > parse empty fields (also confirmed by tech support) !!! I thought it was
> > just my problem but since you bring it to our attention I can confirm
> > that my solution was similar, resorting to replace and textscanning after.
> ...
>
> Oleg, out of curiosity do you happen to know if the behavior mimics C
> scanf() formatting string behavior or is it unique to Matlab? I don't
> know C well enough to have a clue (and am not sure there's a
> working-at-the-moment C compiler handy to give a short test)...
>
> --
Not fluent in C...sorry.

Oleg
 |  Next  |  Last
Pages: 1 2
Prev: Creating a logical mask from a 3D convex hull
Next: About behavior of KeyPressFcn