[Scilab-users] Reading numerical and string vectors from txt-file

[Jens Simon Strom]

mgetl(filename)  is a good alternative. I have used it in a function 
because retrieving numbers comes up frequently for me.

function  cv=!cv(fp,colspan,varargin)//Extracts a numeric Colum Vector from ASCII file
   //Data separator in file:  spaces only, number may differ from line to line to cope with ragged colums
   //fp: full path of ASCII file, type string
   //colspan: Increasing row vector of colum indices, position range of the numbers, type constant
   //         The data colum may be ragged but no other characters may protrude into the colum span.
   //varargin: optional string vector, line numbers to read
   //cv: colum vector of read numbers, type constant
   //CAVEAT: If the file uses separating tabs they should be replaced by blanks. colspan should be enclosed
   //        in at least one leading and one trailing blank in all lines (check in editor).
   //Examples:
   //cv=!cv('file.txt',[12:15]) //reads colums 12 to 15 of all lines
   //cv=!cv('file.txt',[1:6],'[10:$-3]') //reads colums 1 to 6 of 10th to third last line, good for masking headers and footers
   //cv=!cv('file.txt',[1:6],'[1 11 101]') //reads distant lines
   fdescr=mopen(fp,'r');//file descriptor
   txtlines  =  mgetl(fdescr,-1);//string colum vector with all lines
   mclose(fdescr);
   [lhs,rhs]=argn(0)
   if  rhs==3  then
       execstr('txtlines=txtlines('+varargin(1)+')')//reduction to the requested lines
   end
   cv  =   strtod(part(txtlines,colspan))//string to double
endfunction

Jens
---------------------------------------------------------------------------------------------------

Am 23.03.2017 09:51, schrieb Jan Åge Langeland:
>
>
>
> On 23.03.2017 00:25, Jens Simon Strom wrote:
>> The test data I posted had been preprocessed manually by equalizing 
>> the spaces between the colums to a single one. The original output of 
>> the MICA astro software aligns the colums tidily so that the number 
>> of spaces varies along a line and from line to line if the values 
>> vary between 1 and 3 digits. One can fix that with regex. It would be 
>> workflow supporting if the separator input ofcsvRead would interpret 
>> several blanks as equivalent to one or if automation could be 
>> acchieved elsewise.
>> Jens
>> --------------------------------------------------------------------------------------------------------------------------------------------------
>>
> If you have fixed column widths, maybe something like this is better:
>
> Ma=mgetl(filename)
> skiplines=7
> clear Md
> for i=skiplines+1:size(Ma,1)
> Md(i-skiplines,1:22)=strsplit(Ma(i),[4 8 12 17 20 21 24 26 27 29 33 35 
> 36 38 42 45 46 48 53 55 56])'
> end
>
> Jan-Åge
>
>
> _______________________________________________
> users mailing list
> users at lists.scilab.org
> http://lists.scilab.org/mailman/listinfo/users

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.scilab.org/pipermail/users/attachments/20170324/cc88a2b5/attachment.htm>