Menu
▾
▴
[Octave-cvsupdate] SF.net SVN: octave:[7936] trunk/octave-forge/extra/dataframe/inst/@ dataframe
|
From: <cde...@us...> - 2010-11-22 16:48:34
|
Revision: 7936
http://octave.svn.sourceforge.net/octave/?rev=7936&view=rev
Author: cdemills
Date: 2010-11-22 16:48:27 +0000 (Mon, 22 Nov 2010)
Log Message:
-----------
Better handling of missing files, initialisation from scalar/vector, display of complex numbers
Modified Paths:
--------------
trunk/octave-forge/extra/dataframe/inst/@dataframe/dataframe.m
trunk/octave-forge/extra/dataframe/inst/@dataframe/display.m
trunk/octave-forge/extra/dataframe/inst/@dataframe/isfield.m
trunk/octave-forge/extra/dataframe/inst/@dataframe/subsasgn.m
Added Paths:
-----------
trunk/octave-forge/extra/dataframe/inst/@dataframe/sum.m
Property Changed:
----------------
trunk/octave-forge/extra/dataframe/inst/@dataframe/isfield.m
Modified: trunk/octave-forge/extra/dataframe/inst/@dataframe/dataframe.m
===================================================================
--- trunk/octave-forge/extra/dataframe/inst/@dataframe/dataframe.m 2010-11-22 12:16:18 UTC (rev 7935)
+++ trunk/octave-forge/extra/dataframe/inst/@dataframe/dataframe.m 2010-11-22 16:48:27 UTC (rev 7936)
@@ -71,6 +71,7 @@
df._ridx = [];
df._data = cell(0, 0);
df._type = cell(0, 0);
+ df._src = cell(0, 0);
df = class(df, 'dataframe');
return
endif
@@ -147,111 +148,128 @@
if isa(x, 'char') && size(x, 1) < 2,
%# read the data frame from a file
try
- x = load(tilde_expand(x));
+ dummy = tilde_expand(x);
+ x = load(dummy);
+ df._src{end+1, 1} = dummy;
catch
- UTF8_BOM = char([0xEF 0xBB 0xBF]);
+ %# try our own method
+ UTF8_BOM = char([0xEF 0xBB 0xBF]);
unwind_protect
- fid = fopen(tilde_expand(x));
- dummy = fgetl(fid);
- if !strcmp(dummy, UTF8_BOM),
- frewind(fid);
+ dummy = tilde_expand(x);
+ fid = fopen(dummy);
+ if fid != -1,
+ dummy = fgetl(fid);
+ if !strcmp(dummy, UTF8_BOM),
+ frewind(fid);
+ endif
+ in = fscanf(fid, "%c"); %# slurps everything
+ df._src{end+1, 1} = dummy;
+ else
+ in = [];
endif
- in = fscanf(fid, "%c"); %# slurps everything
unwind_protect_cleanup
- fclose(fid);
+ if fid != -1, fclose(fid); endif
end_unwind_protect
- %# explicit list taken from 'man pcrepattern' -- we enclose all
- %# vertical separators in case the underlying regexp engine
- %# doesn't have them all.
- eol = '(\r\n|\n|\v|\f|\r|\x85)';
- %# cut into lines -- include the EOL to have a one-to-one
- %# matching between line numbers. Use a non-greedy match.
- lines = regexp(in, ['.*?' eol], 'match');
- dummy = cellfun(@(x) regexp(x, eol), lines);
- %# remove the EOL character(s)
- lines(1==dummy) = {""};
- %# use a positive lookahead -- eol is not part of the match
- lines(dummy > 1) = cellfun(@(x) regexp(x, ['.*?(?=' eol ')'], \
- 'match'), lines(dummy > 1));
- %# a field either starts at a word boundary, either by + - . for
- %# a numeric data, either by ' for a string.
- %# content = cellfun(@(x) regexp(x, '(\b|[-+\.''])[^,]*(''|\b)', 'match'),\
- %# lines, 'UniformOutput', false); %# extract fields
- content = cellfun(@(x) strsplit(x, sep), lines, \
- 'UniformOutput', false); %# extract fields
- indl = 1; indj = 1; %# disp('line 151 '); keyboard
- if ~isempty(seeked),
+ if !isempty(in),
+ %# explicit list taken from 'man pcrepattern' -- we enclose all
+ %# vertical separators in case the underlying regexp engine
+ %# doesn't have them all.
+ eol = '(\r\n|\n|\v|\f|\r|\x85)';
+ %# cut into lines -- include the EOL to have a one-to-one
+ %# matching between line numbers. Use a non-greedy match.
+ lines = regexp(in, ['.*?' eol], 'match');
+ dummy = cellfun(@(x) regexp(x, eol), lines);
+ %# remove the EOL character(s)
+ lines(1==dummy) = {""};
+ %# use a positive lookahead -- eol is not part of the match
+ lines(dummy > 1) = cellfun(@(x) regexp(x, ['.*?(?=' eol ')'], \
+ 'match'), lines(dummy > 1));
+ %# a field either starts at a word boundary, either by + - . for
+ %# a numeric data, either by ' for a string.
+
+ %# content = cellfun(@(x) regexp(x, '(\b|[-+\.''])[^,]*(''|\b)', 'match'),\
+ %# lines, 'UniformOutput', false); %# extract fields
+ content = cellfun(@(x) strsplit(x, sep), lines, \
+ 'UniformOutput', false); %# extract fields
+ indl = 1; indj = 1; %# disp('line 151 '); keyboard
+ if ~isempty(seeked),
+ while indl <= length(lines),
+ dummy = content{indl};
+ if strcmp(dummy{1}, seeked)
+ break;
+ endif
+ indl = indl + 1;
+ endwhile
+ %# else
+ %# dummy = content{indl};
+ endif
+ x = cell(1+length(lines)-indl, size(dummy, 2));
+ empty_lines = [];
while indl <= length(lines),
dummy = content{indl};
- if strcmp(dummy{1}, seeked)
- break;
+ if all(cellfun('size', dummy, 2) == 0),
+ empty_lines = [empty_lines indj];
+ indl = indl + 1; indj = indj + 1;
+ continue;
endif
- indl = indl + 1;
+ %# does it looks like a comment line ?
+ if regexp(dummy{1}, ['^\s*' char(35)]),
+ empty_lines = [empty_lines indj];
+ indl = indl + 1; indj = indj + 1;
+ continue;
+ endif
+ %# try to convert to float
+ the_line = cellfun(@(x) sscanf(x, "%f"), dummy, \
+ 'UniformOutput', false);
+ for indk = 1: size(the_line, 2),
+ if isempty(the_line{indk}) || any(size(the_line{indk}) > 1),
+ %#if indi > 1 && indk > 1, disp('line 117 '); keyboard; %#endif
+ if unquot,
+ try
+ %# remove quotes and leading space(s)
+ x(indj, indk) = regexp(dummy{indk}, '[^'' ].*[^'']', 'match'){1};
+ catch
+ %# if the previous test fails, try a simpler one
+ in = regexp(dummy{indk}, '[^'' ]+', 'match');
+ if !isempty(in),
+ x(indj, indk) = in{1};
+ %# else
+ %# x(indj, indk) = [];
+ endif
+ end_try_catch
+ else
+ %# no conversion possible, store and remove leading space(s)
+ x(indj, indk) = regexp(dummy{indk}, '[^ ].*', 'match');
+ endif
+ else
+ x(indj, indk) = the_line{indk};
+ endif
+ endfor
+ indl = indl + 1; indj = indj + 1;
endwhile
- %# else
- %# dummy = content{indl};
- endif
- x = cell(1+length(lines)-indl, size(dummy, 2));
- empty_lines = [];
- while indl <= length(lines),
- dummy = content{indl};
- if all(cellfun('size', dummy, 2) == 0),
- empty_lines = [empty_lines indj];
- indl = indl + 1; indj = indj + 1;
- continue;
+ if !isempty(empty_lines),
+ x(empty_lines, :) = [];
endif
- %# does it looks like a comment line ?
- if regexp(dummy{1}, ['^\s*' char(35)]),
- empty_lines = [empty_lines indj];
- indl = indl + 1; indj = indj + 1;
- continue;
+ %# detect empty columns
+ empty_lines = find(0 == sum(cellfun('size', x, 2)));
+ if !isempty(empty_lines),
+ x(:, empty_lines) = [];
endif
- %# try to convert to float
- the_line = cellfun(@(x) sscanf(x, "%f"), dummy, \
- 'UniformOutput', false);
- for indk = 1: size(the_line, 2),
- if isempty(the_line{indk}) || any(size(the_line{indk}) > 1),
- %#if indi > 1 && indk > 1, disp('line 117 '); keyboard; %#endif
- if unquot,
- try
- %# remove quotes and leading space(s)
- x(indj, indk) = regexp(dummy{indk}, '[^'' ].*[^'']', 'match'){1};
- catch
- %# if the previous test fails, try a simpler one
- in = regexp(dummy{indk}, '[^'' ]+', 'match');
- if !isempty(in),
- x(indj, indk) = in{1};
- %# else
- %# x(indj, indk) = [];
- endif
- end_try_catch
- else
- %# no conversion possible, store and remove leading space(s)
- x(indj, indk) = regexp(dummy{indk}, '[^ ].*', 'match');
- endif
- else
- x(indj, indk) = the_line{indk};
- endif
- endfor
- indl = indl + 1; indj = indj + 1;
- endwhile
- if !isempty(empty_lines),
- x(empty_lines, :) = [];
+ clear UTF8_BOM fid in lines indl the_line content empty_lines
endif
- %# detect empty columns
- empty_lines = find(0 == sum(cellfun('size', x, 2)));
- if !isempty(empty_lines),
- x(:, empty_lines) = [];
- endif
- clear UTF8_BOM fid in lines indl the_line content empty_lines
end_try_catch
endif
%# fallback, avoiding a recursive call
idx.type = '()';
- indj = df._cnt(2)+(1:size(x, 2));
-
+ if !isa(x, 'char'),
+ indj = df._cnt(2)+(1:size(x, 2));
+ else
+ %# at this point, reading some filename failed
+ error("dataframe: can't open '%s' for reading data", x);
+ endif;
+
if iscell(x),
if 2 == length(x),
%# use the intermediate value as destination column
@@ -280,7 +298,7 @@
%# allow overwriting of column names
df._over{2}(1, indj) = true;
else
- if !isempty(indj),
+ if !isempty(indj),
if length(df._name{2}) < indj(1) || isempty(df._name{2}(indj)),
[df._name{2}(indj, 1), df._over{2}(1, indj)] ...
= df_colnames(inputname(indi), indj);
@@ -315,14 +333,24 @@
if any([index(base, "=")]),
%# takes the left part as base
x = strsplit(base, "=");
- x = deblank(x{1}); y = false;
- elseif any([index(base, '''')]),
- %# base is most probably a filename
- x = regexp(base, '[^''].*[^'']', 'match'){1}; y = true;
- elseif any([index(base, "(") index(base, ":")]),
- x = 'X'; y = true; %# this is a default value, may be changed
+ x = deblank(x{1});
+ if isvarname(x),
+ y = false;
+ else
+ x = 'X'; y = true;
+ endif
else
- x = base; y = false;
+ %# is base most probably a filename ?
+ x = regexp(base, '''[^''].*[^'']''', 'match');
+ if isempty(x),
+ if isvarname(base),
+ x = base; y = false;
+ else
+ x = 'X'; y = true; %# this is a default value, may be changed
+ endif
+ else
+ x = x{1}; y = true;
+ endif
endif
if numel(num) > 1,
Modified: trunk/octave-forge/extra/dataframe/inst/@dataframe/display.m
===================================================================
--- trunk/octave-forge/extra/dataframe/inst/@dataframe/display.m 2010-11-22 12:16:18 UTC (rev 7935)
+++ trunk/octave-forge/extra/dataframe/inst/@dataframe/display.m 2010-11-22 16:48:27 UTC (rev 7936)
@@ -83,13 +83,12 @@
%# keep only one horizontal space per line
indk = 1; while indk <= size(df._data{indc}, 2),
dummy{3, 2+indi} = disp(df._data{indc}(:, indk));
- tmp_str = char(regexp(dummy{3, 2+indi}, ' \S.*', ...
+ tmp_str = char(regexp(dummy{3, 2+indi}, \
+ '[-+]?[0-9]*\.?[0-9]+([eE][-+]?[0-9]+)?(\s??[-+]\s??[0-9]*\.?[0-9]+([eE][-+]?[0-9]+)?i)?', \
'match', 'dotexceptnewline'));
- if size(tmp_str, 1) < df._cnt(1),
- tmp_str = horzcat...
- (vspace, char(regexp(dummy{3, 2+indi}, '\S.*', ...
- 'match', 'dotexceptnewline')));
- endif
+ tmp_str = horzcat...
+ (vspace, char(regexp(dummy{3, 2+indi}, '\S.*', ...
+ 'match', 'dotexceptnewline')));
dummy{3, 2+indi} = tmp_str;
indi = indi + 1; indk = indk + 1;
endwhile
Modified: trunk/octave-forge/extra/dataframe/inst/@dataframe/isfield.m
===================================================================
--- trunk/octave-forge/extra/dataframe/inst/@dataframe/isfield.m 2010-11-22 12:16:18 UTC (rev 7935)
+++ trunk/octave-forge/extra/dataframe/inst/@dataframe/isfield.m 2010-11-22 16:48:27 UTC (rev 7936)
@@ -32,7 +32,7 @@
%% Suite 330, Boston, MA 02111-1307, USA.
%#
- %# $Id: dataframe.m 7931 2010-11-17 17:07:37Z cdemills $
+ %# $Id$
%#
if !isa(df, 'dataframe'),
Property changes on: trunk/octave-forge/extra/dataframe/inst/@dataframe/isfield.m
___________________________________________________________________
Added: svn:keywords
+ Id
Modified: trunk/octave-forge/extra/dataframe/inst/@dataframe/subsasgn.m
===================================================================
--- trunk/octave-forge/extra/dataframe/inst/@dataframe/subsasgn.m 2010-11-22 12:16:18 UTC (rev 7935)
+++ trunk/octave-forge/extra/dataframe/inst/@dataframe/subsasgn.m 2010-11-22 16:48:27 UTC (rev 7936)
@@ -189,7 +189,7 @@
else
if isvector(RHS),
if 0 == df._cnt(1),
- nrow = length(RHS); %# try to produce row vectors
+ nrow = size(RHS, 1);
else
nrow = df._cnt(1); %# limit to df numbner of rows
endif
@@ -308,7 +308,7 @@
endif
endif
endif
-
+
%# perform row resizing if columns are already filled
if !isempty(indr) && isnumeric(indr),
if max(indr) > df._cnt(1) && size(df._data, 2) == df._cnt(2),
@@ -453,8 +453,9 @@
else
%# ignore 'column' dimension -- force colum vectors -- use a
%# third dim just in case
+ if isempty(S.subs{1}), S.subs{1} = ':'; endif
S.subs(2) = [];
- if length(S.subs) < 2, S.subs(2) = 1; endif
+ if length(S.subs) < 2, S.subs{2} = 1; endif
if length(indc) > 1 && length(RHS) > 1,
%# set a row from a vector
fillfunc = @(x, y) builtin('subsasgn', x, S, RHS(y));
@@ -463,7 +464,11 @@
endif
endif
for indi = 1:length(indc),
- df._data{indc(indi)} = fillfunc(df._data{indc(indi)}, indi);
+ try
+ df._data{indc(indi)} = fillfunc(df._data{indc(indi)}, indi);
+ catch
+ disp('line 470 '); keyboard
+ end_try_catch
# catch
# if ndims(df._data{indc(indi)}) > 2,
# %# upstream forgot to give the third dim
Added: trunk/octave-forge/extra/dataframe/inst/@dataframe/sum.m
===================================================================
--- trunk/octave-forge/extra/dataframe/inst/@dataframe/sum.m (rev 0)
+++ trunk/octave-forge/extra/dataframe/inst/@dataframe/sum.m 2010-11-22 16:48:27 UTC (rev 7936)
@@ -0,0 +1,111 @@
+function resu = sum(df, varargin)
+
+ %# -*- texinfo -*-
+ %# @deftypefn {Function File} {} sum (@var{x})
+ %# @deftypefnx {Function File} {} sum (@var{x}, @var{dim})
+ %# @deftypefnx {Function File} {} sum (@dots{}, 'native')
+ %# @deftypefnx {Function File} {} sum (@dots{}, 'double')
+ %# @deftypefnx {Function File} {} sum (@dots{}, 'extra')
+ %# Sum of elements along dimension @var{dim}. If @var{dim} is
+ %# omitted, it defaults to the first non-singleton dimension.
+ %#
+ %# If the optional argument 'native' is given, then the sum is performed
+ %# in the same type as the original argument, rather than in the default
+ %# double type. For example:
+ %#
+ %# @example
+ %# @group
+ %# sum ([true, true])
+ %# @result{} 2
+ %# sum ([true, true], 'native')
+ %# @result{} true
+ %# @end group
+ %# @end example
+ %#
+ %# On the contrary, if 'double' is given, the sum is performed in double
+ %# precision even for single precision inputs.
+ %#
+ %# For double precision inputs, 'extra' indicates that a more accurate algorithm
+ %# than straightforward summation is to be used. For single precision inputs,
+ %# 'extra' is the same as 'double'. Otherwise, 'extra' has no effect.\n\
+ %# @end deftypefn
+
+ %% Copyright (C) 2009-2010 Pascal Dupuis <Pas...@uc...>
+ %%
+ %% This file is part of Octave.
+ %%
+ %% Octave is free software; you can redistribute it and/or
+ %% modify it under the terms of the GNU General Public
+ %% License as published by the Free Software Foundation;
+ %% either version 2, or (at your option) any later version.
+ %%
+ %% Octave is distributed in the hope that it will be useful,
+ %% but WITHOUT ANY WARRANTY; without even the implied
+ %% warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR
+ %% PURPOSE. See the GNU General Public License for more
+ %% details.
+ %%
+ %% You should have received a copy of the GNU General Public
+ %% License along with Octave; see the file COPYING. If not,
+ %% write to the Free Software Foundation, 59 Temple Place -
+ %% Suite 330, Boston, MA 02111-1307, USA.
+
+ %#
+ %# $Id$
+ %#
+
+
+
+ if !isa(df, 'dataframe'),
+ resu = []; return;
+ endif
+
+ dim = []; type_comp = [];
+
+ indi = 1; while indi <= length(varargin)
+ if isnumeric(varargin{indi}),
+ if !isempty(dim),
+ print_usage();
+ resu = [];
+ return
+ else
+ dim = varargin{indi};
+ endif
+ else
+ if !isempty(type_comp),
+ print_usage();
+ resu = [];
+ return
+ else
+ type_comp = varargin{indi};
+ endif
+ endif
+ endwhile;
+
+ if isempty(dim), dim = 1; endif;
+ if isempty(type_comp), type_comp = 'double'; endif
+
+ %# pre-assignation
+ resu = struct(df);
+
+ switch(dim)
+ case {1},
+ resu._ridx = 1; resu._name{1, 1} = []; resu._over{1, 1} = [];
+ for indi = 1:resu._cnt(2),
+ resu._data{1, indi} = sum(resu._data{1, indi}, dim, type_comp);
+ endfor
+ resu._cnt(1) = 1;
+ case {2},
+ error('Operation not implemented');
+ case {3},
+ for indi = 1:resu._cnt(2),
+ resu._data{1, indi} = sum(resu._data{1, indi}, 2, type_comp)
+ endfor
+ if length(resu._cnt) > 2, resu._cnt = resu._cnt(1:2); endif
+ otherwise
+ error("Invalid dimension %d", dim);
+ endswitch
+
+ resu = dataframe(resu);
+
+endfunction
Property changes on: trunk/octave-forge/extra/dataframe/inst/@dataframe/sum.m
___________________________________________________________________
Added: svn:keywords
+ Id
This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.
|
