view libinterp/corefcn/lookup.cc @ 31605:e88a07dec498 stable

maint: Use macros to begin/end C++ namespaces. * oct-conf-post-public.in.h: Define two macros (OCTAVE_BEGIN_NAMESPACE, OCTAVE_END_NAMESPACE) that can be used to start/end a namespace. * mk-opts.pl, build-env.h, build-env.in.cc, __betainc__.cc, __contourc__.cc, __dsearchn__.cc, __eigs__.cc, __expint__.cc, __ftp__.cc, __gammainc__.cc, __ichol__.cc, __ilu__.cc, __isprimelarge__.cc, __lin_interpn__.cc, __magick_read__.cc, __pchip_deriv__.cc, __qp__.cc, amd.cc, auto-shlib.cc, auto-shlib.h, balance.cc, base-text-renderer.cc, base-text-renderer.h, besselj.cc, bitfcns.cc, bsxfun.cc, c-file-ptr-stream.cc, c-file-ptr-stream.h, call-stack.cc, call-stack.h, ccolamd.cc, cellfun.cc, chol.cc, colamd.cc, colloc.cc, conv2.cc, daspk.cc, dasrt.cc, dassl.cc, data.cc, data.h, debug.cc, defaults.cc, defaults.h, defun-int.h, defun.cc, det.cc, dirfns.cc, display.cc, display.h, dlmread.cc, dmperm.cc, dot.cc, dynamic-ld.cc, dynamic-ld.h, eig.cc, ellipj.cc, environment.cc, environment.h, error.cc, error.h, errwarn.h, event-manager.cc, event-manager.h, event-queue.cc, event-queue.h, fcn-info.cc, fcn-info.h, fft.cc, fft2.cc, fftn.cc, file-io.cc, filter.cc, find.cc, ft-text-renderer.cc, ft-text-renderer.h, gcd.cc, getgrent.cc, getpwent.cc, getrusage.cc, givens.cc, gl-render.cc, gl-render.h, gl2ps-print.cc, gl2ps-print.h, graphics-toolkit.cc, graphics-toolkit.h, graphics.cc, graphics.in.h, gsvd.cc, gtk-manager.cc, gtk-manager.h, hash.cc, help.cc, help.h, hess.cc, hex2num.cc, hook-fcn.cc, hook-fcn.h, input.cc, input.h, interpreter-private.cc, interpreter-private.h, interpreter.cc, interpreter.h, inv.cc, jsondecode.cc, jsonencode.cc, kron.cc, latex-text-renderer.cc, latex-text-renderer.h, load-path.cc, load-path.h, load-save.cc, load-save.h, lookup.cc, ls-ascii-helper.cc, ls-ascii-helper.h, ls-oct-text.cc, ls-utils.cc, ls-utils.h, lsode.cc, lu.cc, mappers.cc, matrix_type.cc, max.cc, mex-private.h, mex.cc, mgorth.cc, nproc.cc, oct-fstrm.cc, oct-fstrm.h, oct-hdf5-types.cc, oct-hdf5-types.h, oct-hist.cc, oct-hist.h, oct-iostrm.cc, oct-iostrm.h, oct-opengl.h, oct-prcstrm.cc, oct-prcstrm.h, oct-procbuf.cc, oct-procbuf.h, oct-process.cc, oct-process.h, oct-stdstrm.h, oct-stream.cc, oct-stream.h, oct-strstrm.cc, oct-strstrm.h, oct-tex-lexer.in.ll, oct-tex-parser.yy, ordqz.cc, ordschur.cc, pager.cc, pager.h, pinv.cc, pow2.cc, pr-flt-fmt.cc, pr-output.cc, procstream.cc, procstream.h, psi.cc, qr.cc, quad.cc, quadcc.cc, qz.cc, rand.cc, rcond.cc, regexp.cc, schur.cc, settings.cc, settings.h, sighandlers.cc, sighandlers.h, sparse-xdiv.cc, sparse-xdiv.h, sparse-xpow.cc, sparse-xpow.h, sparse.cc, spparms.cc, sqrtm.cc, stack-frame.cc, stack-frame.h, stream-euler.cc, strfind.cc, strfns.cc, sub2ind.cc, svd.cc, sylvester.cc, symbfact.cc, syminfo.cc, syminfo.h, symrcm.cc, symrec.cc, symrec.h, symscope.cc, symscope.h, symtab.cc, symtab.h, syscalls.cc, sysdep.cc, sysdep.h, text-engine.cc, text-engine.h, text-renderer.cc, text-renderer.h, time.cc, toplev.cc, tril.cc, tsearch.cc, typecast.cc, url-handle-manager.cc, url-handle-manager.h, urlwrite.cc, utils.cc, utils.h, variables.cc, variables.h, xdiv.cc, xdiv.h, xnorm.cc, xnorm.h, xpow.cc, xpow.h, __delaunayn__.cc, __fltk_uigetfile__.cc, __glpk__.cc, __init_fltk__.cc, __init_gnuplot__.cc, __ode15__.cc, __voronoi__.cc, audiodevinfo.cc, audioread.cc, convhulln.cc, fftw.cc, gzip.cc, mk-build-env-features.sh, mk-builtins.pl, cdef-class.cc, cdef-class.h, cdef-fwd.h, cdef-manager.cc, cdef-manager.h, cdef-method.cc, cdef-method.h, cdef-object.cc, cdef-object.h, cdef-package.cc, cdef-package.h, cdef-property.cc, cdef-property.h, cdef-utils.cc, cdef-utils.h, ov-base.cc, ov-base.h, ov-bool-mat.cc, ov-builtin.h, ov-cell.cc, ov-class.cc, ov-class.h, ov-classdef.cc, ov-classdef.h, ov-complex.cc, ov-fcn-handle.cc, ov-fcn-handle.h, ov-fcn.h, ov-java.cc, ov-java.h, ov-mex-fcn.h, ov-null-mat.cc, ov-oncleanup.cc, ov-struct.cc, ov-typeinfo.cc, ov-typeinfo.h, ov-usr-fcn.cc, ov-usr-fcn.h, ov.cc, ov.h, octave.cc, octave.h, mk-ops.sh, op-b-b.cc, op-b-bm.cc, op-b-sbm.cc, op-bm-b.cc, op-bm-bm.cc, op-bm-sbm.cc, op-cdm-cdm.cc, op-cell.cc, op-chm.cc, op-class.cc, op-cm-cm.cc, op-cm-cs.cc, op-cm-m.cc, op-cm-s.cc, op-cm-scm.cc, op-cm-sm.cc, op-cs-cm.cc, op-cs-cs.cc, op-cs-m.cc, op-cs-s.cc, op-cs-scm.cc, op-cs-sm.cc, op-dm-dm.cc, op-dm-scm.cc, op-dm-sm.cc, op-dm-template.cc, op-dms-template.cc, op-fcdm-fcdm.cc, op-fcm-fcm.cc, op-fcm-fcs.cc, op-fcm-fm.cc, op-fcm-fs.cc, op-fcn.cc, op-fcs-fcm.cc, op-fcs-fcs.cc, op-fcs-fm.cc, op-fcs-fs.cc, op-fdm-fdm.cc, op-fm-fcm.cc, op-fm-fcs.cc, op-fm-fm.cc, op-fm-fs.cc, op-fs-fcm.cc, op-fs-fcs.cc, op-fs-fm.cc, op-fs-fs.cc, op-i16-i16.cc, op-i32-i32.cc, op-i64-i64.cc, op-i8-i8.cc, op-int-concat.cc, op-m-cm.cc, op-m-cs.cc, op-m-m.cc, op-m-s.cc, op-m-scm.cc, op-m-sm.cc, op-mi.cc, op-pm-pm.cc, op-pm-scm.cc, op-pm-sm.cc, op-pm-template.cc, op-range.cc, op-s-cm.cc, op-s-cs.cc, op-s-m.cc, op-s-s.cc, op-s-scm.cc, op-s-sm.cc, op-sbm-b.cc, op-sbm-bm.cc, op-sbm-sbm.cc, op-scm-cm.cc, op-scm-cs.cc, op-scm-m.cc, op-scm-s.cc, op-scm-scm.cc, op-scm-sm.cc, op-sm-cm.cc, op-sm-cs.cc, op-sm-m.cc, op-sm-s.cc, op-sm-scm.cc, op-sm-sm.cc, op-str-m.cc, op-str-s.cc, op-str-str.cc, op-struct.cc, op-ui16-ui16.cc, op-ui32-ui32.cc, op-ui64-ui64.cc, op-ui8-ui8.cc, ops.h, anon-fcn-validator.cc, anon-fcn-validator.h, bp-table.cc, bp-table.h, comment-list.cc, comment-list.h, filepos.h, lex.h, lex.ll, oct-lvalue.cc, oct-lvalue.h, oct-parse.yy, parse.h, profiler.cc, profiler.h, pt-anon-scopes.cc, pt-anon-scopes.h, pt-arg-list.cc, pt-arg-list.h, pt-args-block.cc, pt-args-block.h, pt-array-list.cc, pt-array-list.h, pt-assign.cc, pt-assign.h, pt-binop.cc, pt-binop.h, pt-bp.cc, pt-bp.h, pt-cbinop.cc, pt-cbinop.h, pt-cell.cc, pt-cell.h, pt-check.cc, pt-check.h, pt-classdef.cc, pt-classdef.h, pt-cmd.h, pt-colon.cc, pt-colon.h, pt-const.cc, pt-const.h, pt-decl.cc, pt-decl.h, pt-eval.cc, pt-eval.h, pt-except.cc, pt-except.h, pt-exp.cc, pt-exp.h, pt-fcn-handle.cc, pt-fcn-handle.h, pt-id.cc, pt-id.h, pt-idx.cc, pt-idx.h, pt-jump.h, pt-loop.cc, pt-loop.h, pt-mat.cc, pt-mat.h, pt-misc.cc, pt-misc.h, pt-pr-code.cc, pt-pr-code.h, pt-select.cc, pt-select.h, pt-spmd.cc, pt-spmd.h, pt-stmt.cc, pt-stmt.h, pt-tm-const.cc, pt-tm-const.h, pt-unop.cc, pt-unop.h, pt-vm-eval.cc, pt-walk.cc, pt-walk.h, pt.cc, pt.h, token.cc, token.h, Range.cc, Range.h, idx-vector.cc, idx-vector.h, range-fwd.h, CollocWt.cc, CollocWt.h, aepbalance.cc, aepbalance.h, chol.cc, chol.h, gepbalance.cc, gepbalance.h, gsvd.cc, gsvd.h, hess.cc, hess.h, lo-mappers.cc, lo-mappers.h, lo-specfun.cc, lo-specfun.h, lu.cc, lu.h, oct-convn.cc, oct-convn.h, oct-fftw.cc, oct-fftw.h, oct-norm.cc, oct-norm.h, oct-rand.cc, oct-rand.h, oct-spparms.cc, oct-spparms.h, qr.cc, qr.h, qrp.cc, qrp.h, randgamma.cc, randgamma.h, randmtzig.cc, randmtzig.h, randpoisson.cc, randpoisson.h, schur.cc, schur.h, sparse-chol.cc, sparse-chol.h, sparse-lu.cc, sparse-lu.h, sparse-qr.cc, sparse-qr.h, svd.cc, svd.h, child-list.cc, child-list.h, dir-ops.cc, dir-ops.h, file-ops.cc, file-ops.h, file-stat.cc, file-stat.h, lo-sysdep.cc, lo-sysdep.h, lo-sysinfo.cc, lo-sysinfo.h, mach-info.cc, mach-info.h, oct-env.cc, oct-env.h, oct-group.cc, oct-group.h, oct-password.cc, oct-password.h, oct-syscalls.cc, oct-syscalls.h, oct-time.cc, oct-time.h, oct-uname.cc, oct-uname.h, action-container.cc, action-container.h, base-list.h, cmd-edit.cc, cmd-edit.h, cmd-hist.cc, cmd-hist.h, f77-fcn.h, file-info.cc, file-info.h, lo-array-errwarn.cc, lo-array-errwarn.h, lo-hash.cc, lo-hash.h, lo-ieee.h, lo-regexp.cc, lo-regexp.h, lo-utils.cc, lo-utils.h, oct-base64.cc, oct-base64.h, oct-glob.cc, oct-glob.h, oct-inttypes.h, oct-mutex.cc, oct-mutex.h, oct-refcount.h, oct-shlib.cc, oct-shlib.h, oct-sparse.cc, oct-sparse.h, oct-string.h, octave-preserve-stream-state.h, pathsearch.cc, pathsearch.h, quit.cc, quit.h, unwind-prot.cc, unwind-prot.h, url-transfer.cc, url-transfer.h : Use new macros to begin/end C++ namespaces.
author Rik <rik@octave.org>
date Thu, 01 Dec 2022 14:23:45 -0800
parents 796f54d4ddbf
children aac27ad79be6
line wrap: on
line source

////////////////////////////////////////////////////////////////////////
//
// Copyright (C) 2008-2022 The Octave Project Developers
//
// See the file COPYRIGHT.md in the top-level directory of this
// distribution or <https://octave.org/copyright/>.
//
// 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 3 of the License, 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, see
// <https://www.gnu.org/licenses/>.
//
////////////////////////////////////////////////////////////////////////

#if defined (HAVE_CONFIG_H)
#  include "config.h"
#endif

#include <cctype>
#include <functional>
#include <algorithm>

#include "dNDArray.h"
#include "CNDArray.h"

#include "Cell.h"
#include "defun.h"
#include "error.h"
#include "errwarn.h"
#include "ovl.h"
#include "ov.h"

OCTAVE_BEGIN_NAMESPACE(octave)

static
bool
contains_char (const std::string& str, char c)
{
  return (str.find (c) != std::string::npos
          || str.find (std::toupper (c)) != std::string::npos);
}

template <typename T>
inline sortmode
get_sort_mode (const Array<T>& array,
               typename octave_sort<T>::compare_fcn_type desc_comp
               = octave_sort<T>::descending_compare)
{
  octave_idx_type n = array.numel ();
  if (n > 1 && desc_comp (array (0), array (n-1)))
    return DESCENDING;
  else
    return ASCENDING;
}

// FIXME: perhaps there should be octave_value::lookup?
// The question is, how should it behave w.r.t. the second argument's type.
// We'd need a dispatch on two arguments.  Hmmm...

#define INT_ARRAY_LOOKUP(TYPE)                                  \
  (table.is_ ## TYPE ## _type () && y.is_ ## TYPE ## _type ())  \
    retval = do_numeric_lookup (table.TYPE ## _array_value (),  \
                                y.TYPE ## _array_value (),      \
                                left_inf, right_inf,            \
                                match_idx, match_bool);
template <typename ArrayT>
static octave_value
do_numeric_lookup (const ArrayT& array, const ArrayT& values,
                   bool left_inf, bool right_inf,
                   bool match_idx, bool match_bool)
{
  octave_value retval;

  Array<octave_idx_type> idx = array.lookup (values);
  octave_idx_type n = array.numel ();
  octave_idx_type nval = values.numel ();

  // Post-process.
  if (match_bool)
    {
      boolNDArray match (idx.dims ());
      for (octave_idx_type i = 0; i < nval; i++)
        {
          octave_idx_type j = idx.xelem (i);
          match.xelem (i) = j != 0 && values(i) == array(j-1);
        }

      retval = match;
    }
  else if (match_idx || left_inf || right_inf)
    {
      if (match_idx)
        {
          NDArray ridx (idx.dims ());

          for (octave_idx_type i = 0; i < nval; i++)
            {
              octave_idx_type j = idx.xelem (i);
              ridx.xelem (i) = (j != 0 && values(i) == array(j-1)) ? j : 0;
            }

          retval = ridx;
        }
      else if (left_inf && right_inf)
        {
          // Results in valid indices.  Optimize using lazy index.
          octave_idx_type zero = 0;
          for (octave_idx_type i = 0; i < nval; i++)
            {
              octave_idx_type j = idx.xelem (i) - 1;
              idx.xelem (i) = std::max (zero, std::min (j, n-2));
            }

          retval = idx_vector (idx);
        }
      else if (left_inf)
        {
          // Results in valid indices.  Optimize using lazy index.
          octave_idx_type zero = 0;
          for (octave_idx_type i = 0; i < nval; i++)
            {
              octave_idx_type j = idx.xelem (i) - 1;
              idx.xelem (i) = std::max (zero, j);
            }

          retval = idx_vector (idx);
        }
      else if (right_inf)
        {
          NDArray ridx (idx.dims ());

          for (octave_idx_type i = 0; i < nval; i++)
            {
              octave_idx_type j = idx.xelem (i);
              ridx.xelem (i) = std::min (j, n-1);
            }

          retval = ridx;
        }
    }
  else
    retval = idx;

  return retval;
}

DEFUN (lookup, args, ,
       doc: /* -*- texinfo -*-
@deftypefn  {} {@var{idx} =} lookup (@var{table}, @var{y})
@deftypefnx {} {@var{idx} =} lookup (@var{table}, @var{y}, @var{opt})
Lookup values in a @strong{sorted} table.

This function is usually used as a prelude to interpolation.

If table is increasing, of length N and @code{idx = lookup (table, y)}, then
@code{table(idx(i)) <= y(i) < table(idx(i+1))} for all @code{y(i)} within the
table.  If @code{y(i) < table(1)} then @code{idx(i)} is 0.  If
@code{y(i) >= table(end)} or @code{isnan (y(i))} then @code{idx(i)} is N.

If the table is decreasing, then the tests are reversed.  For non-strictly
monotonic tables, empty intervals are always skipped.  The result is undefined
if @var{table} is not monotonic, or if @var{table} contains a NaN.

The complexity of the lookup is O(M*log(N)) where M is the size of @var{y}.
In the special case when @var{y} is also sorted, the complexity is
O(min (M*log(N), M+N)).

@var{table} and @var{y} can also be cell arrays of strings (or @var{y} can be a
single string).  In this case, string lookup is performed using lexicographical
comparison.

If @var{opts} is specified, it must be a string with letters indicating
additional options.

@table @code
@item m
Match.  @code{table(idx(i)) == y(i)} if @code{y(i)} occurs in table;
otherwise, @code{idx(i)} is zero.

@item b
Boolean.  @code{idx(i)} is a logical 1 or 0, indicating whether @code{y(i)}
is contained in table or not.

@item l
Left.  For numeric lookups the leftmost subinterval shall be extended to
minus infinity (i.e., all indices at least 1).

@item r
Right.  For numeric lookups the rightmost subinterval shall be extended to
infinity (i.e., all indices at most N-1).
@end table

@strong{Note}: If @var{table} is not sorted the results from @code{lookup}
will be unpredictable.
@end deftypefn */)
{
  int nargin = args.length ();

  if (nargin < 2 || nargin > 3)
    print_usage ();

  octave_value table = args(0);
  octave_value y = args(1);
  if (table.ndims () > 2 || (table.columns () > 1 && table.rows () > 1))
    warning ("lookup: table is not a vector");

  octave_value retval;

  bool num_case = ((table.isnumeric () && y.isnumeric ())
                   || (table.is_char_matrix () && y.is_char_matrix ()));
  bool str_case = table.iscellstr () && (y.is_string () || y.iscellstr ());
  bool left_inf = false;
  bool right_inf = false;
  bool match_idx = false;
  bool match_bool = false;

  if (nargin == 3)
    {
      std::string opt = args(2).xstring_value ("lookup: OPT must be a string");
      left_inf = contains_char (opt, 'l');
      right_inf = contains_char (opt, 'r');
      match_idx = contains_char (opt, 'm');
      match_bool = contains_char (opt, 'b');
      if (opt.find_first_not_of ("lrmb") != std::string::npos)
        error ("lookup: unrecognized option: %c",
               opt[opt.find_first_not_of ("lrmb")]);
    }

  if ((match_idx || match_bool) && (left_inf || right_inf))
    error ("lookup: m, b cannot be specified with l or r");
  else if (match_idx && match_bool)
    error ("lookup: only one of m or b can be specified");
  else if (str_case && (left_inf || right_inf))
    error ("lookup: l, r are not recognized for string lookups");

  if (num_case)
    {
      // In the case of a complex array, absolute values will be used for
      // compatibility (though it's not too meaningful).
      if (table.iscomplex ())
        table = table.abs ();

      if (y.iscomplex ())
        y = y.abs ();

      Array<octave_idx_type> idx;

      // PS: I learned this from data.cc
      if INT_ARRAY_LOOKUP (int8)
      else if INT_ARRAY_LOOKUP (int16)
      else if INT_ARRAY_LOOKUP (int32)
      else if INT_ARRAY_LOOKUP (int64)
      else if INT_ARRAY_LOOKUP (uint8)
      else if INT_ARRAY_LOOKUP (uint16)
      else if INT_ARRAY_LOOKUP (uint32)
      else if INT_ARRAY_LOOKUP (uint64)
      else if (table.is_char_matrix () && y.is_char_matrix ())
        retval = do_numeric_lookup (table.char_array_value (),
                                    y.char_array_value (),
                                    left_inf, right_inf,
                                    match_idx, match_bool);
      else if (table.is_single_type () || y.is_single_type ())
        retval = do_numeric_lookup (table.float_array_value (),
                                    y.float_array_value (),
                                    left_inf, right_inf,
                                    match_idx, match_bool);
      else
        retval = do_numeric_lookup (table.array_value (),
                                    y.array_value (),
                                    left_inf, right_inf,
                                    match_idx, match_bool);
    }
  else if (str_case)
    {
      Array<std::string> str_table = table.cellstr_value ();
      Array<std::string> str_y (dim_vector (1, 1));

      if (y.iscellstr ())
        str_y = y.cellstr_value ();
      else
        str_y(0) = y.string_value ();

      Array<octave_idx_type> idx = str_table.lookup (str_y);
      octave_idx_type nval = str_y.numel ();

      // Post-process.
      if (match_bool)
        {
          boolNDArray match (idx.dims ());
          for (octave_idx_type i = 0; i < nval; i++)
            {
              octave_idx_type j = idx.xelem (i);
              match.xelem (i) = j != 0 && str_y(i) == str_table(j-1);
            }

          retval = match;
        }
      else if (match_idx)
        {
          NDArray ridx (idx.dims ());
          for (octave_idx_type i = 0; i < nval; i++)
            {
              octave_idx_type j = idx.xelem (i);
              ridx.xelem (i) = (j != 0 && str_y(i) == str_table(j-1) ? j : 0);
            }

          retval = ridx;
        }
      else
        retval = idx;
    }
  else
    print_usage ();

  return retval;
}

/*
%!assert (lookup (1:3, 0.5), 0)     # value before table
%!assert (lookup (1:3, 3.5), 3)     # value after table error
%!assert (lookup (1:3, 1.5), 1)     # value within table error
%!assert (lookup (1:3, [3,2,1]), [3,2,1])
%!assert (lookup ([1:4]', [1.2, 3.5]'), [1, 3]')
%!assert (lookup ([1:4], [1.2, 3.5]'), [1, 3]')
%!assert (lookup ([1:4]', [1.2, 3.5]), [1, 3])
%!assert (lookup ([1:4], [1.2, 3.5]), [1, 3])
%!assert (lookup (1:3, [3, 2, 1]), [3, 2, 1])
%!assert (lookup ([3:-1:1], [3.5, 3, 1.2, 2.5, 2.5]), [0, 1, 2, 1, 1])
%!assert (isempty (lookup ([1:3], [])))
%!assert (isempty (lookup ([1:3]', [])))
%!assert (lookup (1:3, [1, 2; 3, 0.5]), [1, 2; 3, 0])
%!assert (lookup (1:4, [1, 1.2; 3, 2.5], "m"), [1, 0; 3, 0])
%!assert (lookup (4:-1:1, [1, 1.2; 3, 2.5], "m"), [4, 0; 2, 0])
%!assert (lookup (1:4, [1, 1.2; 3, 2.5], "b"), logical ([1, 0; 3, 0]))
%!assert (lookup (4:-1:1, [1, 1.2; 3, 2.5], "b"), logical ([4, 0; 2, 0]))
%!
%!assert (lookup ({"apple","lemon","orange"}, {"banana","kiwi"; "ananas","mango"}), [1,1;0,2])
%!assert (lookup ({"apple","lemon","orange"}, "potato"), 3)
%!assert (lookup ({"orange","lemon","apple"}, "potato"), 0)
*/

OCTAVE_END_NAMESPACE(octave)