File: rdcols.i

package info (click to toggle)
yorick-yutils 1.3.0-2
  • links: PTS
  • area: main
  • in suites: lenny
  • size: 604 kB
  • ctags: 17
  • sloc: makefile: 105; python: 12
file content (305 lines) | stat: -rw-r--r-- 10,279 bytes parent folder | download | duplicates (2)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
/*
 * rdcols.i -- $Id: rdcols.i,v 1.1.1.1 2007/12/11 23:55:12 frigaut Exp $
 * routines to assist in reading columns from ascii files
 * 
 * Author: David Munro
 * Written 2002
 *
 * Copyright (c) 2002, David Munro
 *
 * This program 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 of the License,  or  (at your
 * option) any later version.
 *
 * This program 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 (to receive a  copy  of  the GNU
 * General Public License, write to the Free Software Foundation, Inc., 675
 * Mass Ave, Cambridge, MA 02139, USA).
 */

func rdcols(f, ncols, width=, delim=, type=, missing=, marker=,
            comment=, nskip=, nlines=)
/* DOCUMENT cols = rdcols(f)
 *       or cols = rdcols(f, ncols)
 *          eq_nocopy, col1, *cols(1)
 *          eq_nocopy, col2, *cols(2)
 *          ...
 *          eq_nocopy, colN, *cols(ncols)
 *
 *  cracks ascii file F into NCOLS columns, returning an array of
 *  pointers to the columns.  A "column" is defined by either its
 *  width in characters, or by the appearance of a delimiting
 *  character (such as a blank, a comma, or a tab).  Multiple
 *  delimiting characters may either be skipped (as usual if the
 *  character is a blank), or may represent a number of empty
 *  columns (as for reading a typical text file exported from a
 *  spreadsheet or database program).
 *
 *  rdcols returns an array of pointers COLS, with *COLS(i) the
 *  contents of the i-th column.  The NCOLS parameter may be omitted
 *  in which case rdcols will guess the number of columns.
 *
 *  F may be either a file handle or a filename.
 *
 *  nskip=m       skip the first M lines of the file (M=0 by default)
 *  nlines=n      read only the first M+N lines of the file
 *  comment=text  interpret lines beginning with TEXT (possibly preceded
 *                by blanks) as comments (blank lines are always comments)
 *  width=w       column width, or list of widths (w=0 means any width)
 *  delim=d       column delimiters (as for strtok), or list of delimiters
 *  type=t        column type, or list of types
 *    t=0 means guess, t=1 means string, t=2 means integer, t=3 means real,
 *    t=4 means either integer or real
 *  marker=k      marker column delimiters (as for strtok)
 *  missing=m     missing value, or list of values
 *                for empty numeric columns
 *
 *  The width=, delim=, type=, and missing= keywords may either be scalar
 *  values to apply to every column, or lists to apply to successive
 *  columns.  The marker= keyword is similar to the delim= keyword,
 *  except that multiple consecutive occurrences indicate empty
 *  columns.  If any list is shorter than NCOLS, its final value
 *  applies to all remaining columns.  If any list is longer than
 *  NCOLS, its trailing values are silently ignored.
 *
 *  By default, width=0.  Non-zero width is very tricky, since a
 *  column begins wherever the previous column ended; if the previous
 *  column had width=0, its delimiter is treated as a part of it, so
 *  it is not included in the width count.  Consequently, the width=
 *  keyword is best suited to situations in which only the final
 *  width value is 0.
 *
 *  By default, marker=[].  If present, then every column is delimited
 *  by a character in marker, and every occurrence of such a character
 *  indicates a new column.  Use marker="\t" to read tab-delimited text
 *  files exported from a spreadsheet or database program.  Non-zero
 *  width= values supersede marker=, but delim= is ignored if marker
 *  is present.
 *
 *  By default, delim=" \t",  and type=0, so that blanks and tabs delimit
 *  columns, and rdcols guesses whether a column contains string or
 *  numeric data.  A numeric column ends at the first character that
 *  cannot be interpreted as part of the number, whether or not that is
 *  a character in delim.  If you use type= to force a numeric column,
 *  empty rows or rows which are not numbers get the value 0.0 by
 *  default, or the value specified by the missing= keyword.  Missing
 *  values in string columns always get the value "".
 *
 *  If you need finer control of the returned types than provided by
 *  the type= keyword, see rdconvert.
 *
 * SEE ALSO: read, rdline, rdconvert
 */
{
  if (structof(f) == string) f = open(f);

  if (is_void(ncols)) nc = 1;
  else if (ncols < 1) return [];
  else nc = ncols;
  result = array(pointer, nc);

  if (!is_void(nskip) && nskip>0) rdline, f, nskip;
  arb = is_void(nlines);
  if (arb) nlines = 10000;
  if (nlines < 1) return result;

  /* ingest entire file as string array
   * note each read grabs twice as many lins as previous,
   * so grow operation scales as nlines*log(nlines), which is ok */
  lines = _rc_rdline(f, nlines);
  if (arb) {
    for (l=lines ; ; nlines*=2) {
      l = _rc_rdline(f, nlines);
      grow, lines, l;
      if (numberof(l) < nlines) break;
    }
  }
  if (!numberof(lines)) return result;

  /* remove blank and comment lines */
  tok = strtok(lines)(1,);
  mask = !tok;
  if (is_void(comment) && is_void(marker)) comment = "#";
  if (comment && strlen(comment)) {
    comment = strtok(comment)(1);
    mask |= strpart(tok, 1:strlen(comment))==comment;
  }
  if (anyof(mask)) {
    list = where(!mask);
    if (!numberof(list)) return result;
    lines = lines(list);
  }

  if (is_void(width)) width = 0;
  else width = long(width);
  nw = numberof(width);

  if (is_void(type)) type = 0;
  else type = long(type);
  nt = numberof(type);

  if (is_void(missing)) missing = 0.0;
  else missing = double(missing);
  nm = numberof(missing);

  if (is_void(delim)) delim = is_void(marker)? " \t" : string(0);
  nk = 0;
  if (!is_void(marker)) {
    if (strlen(marker) > 0) {
      nk = 1;
      /* valid is a character not in marker */
      valid = strtok(string(&char(indgen(255))), marker)(1);
      valid = strpart(valid,1:1);
    } else {
      marker = [];
    }
  }
  nd = numberof(delim);

  for (i=1 ; ; i++) {
    w = width(min(i,nw));
    d = delim(min(i,nd));
    if (!strlen(d)) d = string(0);
    t = type(min(i,nt));
    m = missing(min(i,nm));
    if (w > 0) {         /* column delimited by its width */
      lines = strpart(lines, w+1:0);
      col = _rc_convert(strpart(lines, 1:w), t, m);
    } else if (nk) {     /* column delimited by marker= */
      col = strtok(valid+lines, marker);
      lines = col(2,);
      col = _rc_convert(strpart(col(1,),2:0), t, m);
    } else {             /* column delimited by delim= */
      col = _rc_convert(lines, t, m, d, lines);
    }
    result(i) = &col;
    if (!is_void(ncols) && i==ncols) break;
    if (noneof(strlen(lines))) break;
    if (i == nc) {
      grow, result, array(pointer, nc);
      nc *= 2;
    }
  }
  if (is_void(ncols)) result = result(1:i);

  return result;
}

func rdconvert(cols, ..)
/* DOCUMENT rdconvert, cols, type1, type2, type3, ...
 *       or cols = rdconvert(cols, type1, type2, type3, ...)
 *
 *  converts the data types of the COLS, so that COLS(1) becomes
 *  TYPE1, COLS(2) TYPE2, COLS(3) TYPE3, and so on.  The COLS
 *  array is an array of pointers as returned by rdcols.  If any
 *  of the TYPEi is nil, that column is not altered.  This function
 *  is only useful if you are not satisfied with either the long
 *  or double types returned by rdcols, e.g.-
 *     cols = rdcols("datafile", 5);
 *     rdconvert, cols, , short, float, , int;
 *
 * SEE ALSO: rdcols
 */
{
  for (i=1 ; more_args() ; i++) {
    type = next_arg();
    if (!is_void(type)) cols(i) = &type(*cols(i));
  }
  return cols;
}

func _rc_rdline(f, nlines)
{  /* read at most nlines lines from f */
  lines = rdline(f, nlines);
  n = sum(!lines);
  if (n) {
    nlines -= n;
    lines = nlines? lines(1:nlines) : [];
  }
  return lines;
}

func _rc_guess(lines)
{
  tok = strtok(lines, " \t")(1,);
  t = strpart(tok, 1:1);
  list = where((t=="-") | (t=="+"));
  if (numberof(list)) {
    tok(list) = x = strpart(tok(list), 2:0);
    t(list) = strpart(x, 1:1);
  }
  list = where(t==".");
  if (numberof(list))
    t(list) = strpart(tok(list), 2:2);
  guess = ((t>="0") & (t<="9"));
  list = where(guess);
  if (numberof(list)) {
    tok = tok(list);
    t = strpart(strtok(tok, "0123456789")(1,), 1:1)
    r = (t==".");
    maybe = where((t=="e") | (t=="E") | (t=="d") | (t=="D"));
    if (numberof(maybe)) {
      u = strtok(tok(maybe), "eEdD")(2,);
      t = strpart(u, 1:1);
      l = where((t=="-") | (t=="+"));
      if (numberof(l)) t(l) = strpart(u(l), 2:2);
      r(maybe) = ((t>="0") & (t<="9"));
    }
    guess(list) += r;
  }
  /* 1 where string, 2 where integer, 3 where real */
  return guess+1;
}

func _rc_convert(lines, type, missing, delim, &remains)
{
  if (type != 1) {
    guess = _rc_guess(lines);
    if (type == 0) {
      /* if more than 80% of a column is numbers, guess it is numbers */
      if (sum(guess==1)>numberof(lines)/5) type = 1;
      else if (anyof(guess==3)) type = 3;
      else type = 2;
    } else if (type!=2 && type!=3) {
      if (anyof(guess==3)) type = 3;
      else type = 2;
    }
  }

  if (type == 1) {
    if (delim) {
      lines = strtok(lines, delim);
      remains = lines(2,);
      lines = lines(1,);
    } else {
      remains = array(string, numberof(lines));
    }
    return lines;
  }

  /* change non-numeric lines into numeric ones */
  list = where(guess==1);
  if (numberof(list)) {
    vals = lines(list);
    lines(list) = "0"+vals;
    if (delim) rems = strtok(vals, delim)(2,);
  }
  /* make sure all lines have at least one non-numeric char at end */
  lines += "?";

  vals = array(0.0, numberof(lines));
  remains = array(string, numberof(lines));
  if (sread(lines,format="%le%[^\n]",vals,remains) != 2*numberof(lines))
    error, "impossible error during rdcols()";
  remains = strpart(remains, 1:-1);
  if (numberof(list)) {
    vals(list) = missing;
    if (delim) remains(list) = rems;
  }

  if (type == 2) vals = long(vals);
  return vals;
}