| Re: Reading and Plotting big txt. File [message #55097] |
Wed, 01 August 2007 09:43  |
Conor
Messages: 138
Registered: February 2007
|
Senior Member |
|
|
On Aug 1, 12:31 pm, "incognito.me" <incognito...@gmx.de> wrote:
> On 1 Aug., 18:15, Conor <cmanc...@gmail.com> wrote:
>
>
>
>> On Aug 1, 10:49 am, "incognito.me" <incognito...@gmx.de> wrote:
>
>>> On 1 Aug., 14:44, Conor <cmanc...@gmail.com> wrote:
>
>>>> On Aug 1, 6:25 am, greg.a...@googlemail.com wrote:
>
>>>> > On Aug 1, 11:33 am, "incognito.me" <incognito...@gmx.de> wrote:
>
>>>> > > I'm trying to read and plot (surface) a very big text (.txt) file
>>>> > > (1020, 1024) with a 5 line string Header in IDL. My file looks like a
>>>> > > circle made of numbers!!!. That means in some lines and colums there
>>>> > > are no numbers only blanks!!!for example my file contains integers
>>>> > > between rows 633 and 390 and between columns 650 and 406.At the left
>>>> > > side of the file, there are the numbers of rows (1023,1022,1021,....0)
>>>> > > my code should not read, but it does. And I also notice, that my code
>>>> > > don't begin to read where the data starts!!By running the code I have
>>>> > > the following error message: READF: End of file encountered. Unit: 1.
>>>> > > Can someone help me?
>>>> > > This is how my code looks like
>>>> > > pro readfile, filename
>
>>>> > > ; file=strupcase(filename)
>>>> > > rows=file_lines(file)
>>>> > > ;open the file and read the five line header.
>>>> > > openr,1,file
>>>> > > header=strarr(5)
>>>> > > readf,1,header
>>>> > > ; Find the number of columns in the file
>>>> > > cols=fix(strmid(header(3),14,4))
>>>> > > ; Number of rows of the data
>>>> > > rows_data=rows-n_elements(header)
>
>>>> > > ;Create a big array to hold the data
>>>> > > data=intarr (cols, rows_data)
>>>> > > ; All blanks should be replaced by zero
>>>> > > data[where(data eq ' ')]=0
>>>> > > ; A small array to read a line
>>>> > > s=intarr(cols)
>>>> > > n=0
>>>> > > while (~ eof(1) and (n lt rows_data -1 )) do begin
>>>> > > ; Read a line of data
>>>> > > readf,1,s
>>>> > > ; Store it in data
>>>> > > data[*,n]=s
>>>> > > n=n+1
>>>> > > end
>>>> > > data=data[*,0:n-1]
>
>>>> > > CLOSE,1
>>>> > > Shade_surf, data
>>>> > > end
>
>>>> > > thanks
>
>>>> > > incognito
>
>>>> > I'm suspicious of the line converting blanks to zeros before you've
>>>> > even read them. I don't think the blanks will come out the way you're
>>>> > expecting, anyway. I'd suggest you write a program to correctly read
>>>> > your first line of data before you go for the whole thing.
>
>>>> > Greg
>
>>>> For starters, I'm not sure why you are converting blanks to zeroes
>>>> there at all. As far as I can tell, you haven't even initialized any
>>>> data yet. It seems like you are trying to convert blanks to zeros on
>>>> an integer array which is already filled with zeroes anyway. When I
>>>> tried to do that, I got this error:
>
>>>> % Type conversion error: Unable to convert given STRING to Integer.
>
>>>> Which isn't a fatal error, so your code would still run but the line
>>>> 'data[where(data eq ' ')]=0' wouldn't actually do anything. As for
>>>> the rest of your problem, I think what you need is a format
>>>> statement. I believe what is happening is that because you haven't
>>>> included an explicit format statement (telling it how many columns are
>>>> on each line) it simply reads in entries until it fills up a row in
>>>> your data array. For instance, look at this file:
>
>>>> 12 34 698 934
>>>> 16 18
>>>> 17 20 13
>>>> 14 23 234 123
>
>>>> being read by this pseudo-code:
>
>>>> readf,lun,file,/get_lun
>>>> data = intarr(4)
>>>> readf,lun,data
>>>> print,data
>>>> ; 12 34 698 934
>>>> readf,lun,data
>>>> print,data
>>>> ; 16 13 17 20
>>>> readf,lun,data
>>>> print,data
>>>> ; 14 23 234 123
>>>> readf,lun,data
>>>> % READF: End of file encountered. Unit: 100, File: test
>
>>>> See, because you have no format specified, each readf keeps reading
>>>> data in until the data array is filled. You are assuming that readf
>>>> reads one line at a time, but that's not happening, which is why your
>>>> data isn't where it's supposed to be. Also, because it is reading
>>>> faster than one line at a time, you are reading to the end of the file
>>>> before you call readf (rows_data) times, and then you get the EOF
>>>> error. The solution is to give it a format:
>
>>>> IDL> openr,lun,'test',/get_lun
>>>> IDL> format = '(i3, 1x, i3, 1x, i3, 1x, i3)'
>>>> IDL> readf,lun,test,format=format
>>>> IDL> print,test
>>>> 12 34 698 934
>>>> IDL> readf,lun,test,format=format
>>>> IDL> print,test
>>>> 16 0 0 18
>>>> IDL> readf,lun,test,format=format
>>>> IDL> print,test
>>>> 17 20 0 13
>>>> IDL> readf,lun,test,format=format
>>>> IDL> print,test
>>>> 14 23 234 123- Zitierten Text ausblenden -
>
>>>> - Zitierten Text anzeigen -
>
>>> Hi Conor,
>
>>> Thanks for your suggestions!I muss agree,to fill the blanks with
>>> zeroes was not so cute!!I have to read how one uses the keyword format
>>> with readf again,because I should confest I haven't unsterstood
>>> yet.Could you please give me a hint?
>>> Thanks a lot,
>>> Kind regards
>>> C.
>
>> Unfortunately, I'm not so great with format statements, I don't use
>> them so much, and I've never used them for reading files. The general
>> idea for reading floats is that you specify the total number of
>> characters to read, and how many numbers come after the decimal
>> place. So, for instance the number:
>
>> 123.456789
>
>> would be specified by the statement:
>
>> (f10.6)
>
>> There are ten characters that must be read (9 digits, plus the decimal
>> point) and there are 6 digits after the period. For spaces you use
>> '1x' (or '2x' for two spaces, etc...). So for instance the line:
>
>> 134.367 123.45 123.92
>
>> would be specified by:
>
>> (f7.3, 1x, f6.2, 1x, f6.2)
>
>> Also, you can specify that IDL should "repeat" a format statement.
>> For instance, you could also represent the last one with:
>
>> (f7.3, 2(1x, f6.2) )
>
>> This last part is very important to you because you won't want to
>> write out the format statement for all 1000 of your columns. In fact,
>> IDL won't let you specify that many anyway. With any luck, all the
>> columns have the same fixed width (or at least a repeating pattern) so
>> you can do something like this:
>
>> (f10.5, 999(1x, f12.1) )
>
>> Exactly how it will work I don't know. You might just have to play
>> around with it. As I said, I'm not terribly familiar with format
>> statements myself, so this might not be the best way to do it. Maybe
>> someone else has some suggestions?- Zitierten Text ausblenden -
>
>> - Zitierten Text anzeigen -
>
> Hi Conor,
>
> I'm now better unterstanding how the format statement works.I will
> jetzt
> managed to understand how it works with negative integers.I think,it
> won't
> be so different.Thanks a lot for the hint.It was very helpfull.
> Kind regards,
> C.
Negative integers are pretty similar.
-1234
would be:
(i5)
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
