Provided by: tcllib_1.15-dfsg-2_all bug

NAME

       struct::matrix - Create and manipulate matrix objects

SYNOPSIS

       package require Tcl  8.2

       package require struct::matrix  ?2.0.1?

       ::struct::matrix ?matrixName? ?=|:=|as|deserialize source?

       matrixName option ?arg arg ...?

       matrixName = sourcematrix

       matrixName --> destmatrix

       matrixName add column ?values?

       matrixName add row ?values?

       matrixName add columns n

       matrixName add rows n

       matrixName cells

       matrixName cellsize column row

       matrixName columns

       matrixName columnwidth column

       matrixName delete column column

       matrixName delete columns n

       matrixName delete row row

       matrixName delete rows n

       matrixName deserialize serialization

       matrixName destroy

       matrixName format 2string ?report?

       matrixName format 2chan ??report? channel?

       matrixName get cell column row

       matrixName get column column

       matrixName get rect column_tl row_tl column_br row_br

       matrixName get row row

       matrixName insert column column ?values?

       matrixName insert row row ?values?

       matrixName link ?-transpose? arrayvar

       matrixName links

       matrixName rowheight row

       matrixName rows

       matrixName search ?-nocase? ?-exact|-glob|-regexp? all pattern

       matrixName search ?-nocase? ?-exact|-glob|-regexp? column column pattern

       matrixName search ?-nocase? ?-exact|-glob|-regexp? row row pattern

       matrixName search ?-nocase? ?-exact|-glob|-regexp? rect column_tl row_tl column_br row_br pattern

       matrixName serialize ?column_tl row_tl column_br row_br?

       matrixName set cell column row value

       matrixName set column column values

       matrixName set rect column row values

       matrixName set row row values

       matrixName sort columns ?-increasing|-decreasing? row

       matrixName sort rows ?-increasing|-decreasing? column

       matrixName swap columns column_a column_b

       matrixName swap rows row_a row_b

       matrixName transpose

       matrixName unlink arrayvar

_________________________________________________________________

DESCRIPTION

       A  matrix  is  a  rectangular collection of cells, i.e. organized in rows and columns. Each cell contains
       exactly one value of arbitrary form. The cells in the matrix are addressed by pairs of  integer  numbers,
       with  the  first (left) number in the pair specifying the column and the second (right) number specifying
       the row the cell is in. These indices are counted from 0 upward. The special non-numeric index end refers
       to  the  last  row  or column in the matrix, depending on the context. Indices of the form end-number are
       counted from the end of the row or column, like they are for standard Tcl lists. Trying  to  access  non-
       existing cells causes an error.

       The  matrices  here  are created empty, i.e. they have neither rows nor columns. The user then has to add
       rows and columns as needed by his application. A specialty of this structure is the ability to export  an
       array-view  onto  its  contents.  Such  can  be used by tkTable, for example, to link the matrix into the
       display.

       The main command of the package is:

       ::struct::matrix ?matrixName? ?=|:=|as|deserialize source?
              The command creates a new matrix object with an  associated  global  Tcl  command  whose  name  is
              matrixName.   This  command  may  be  used to invoke various operations on the matrix.  It has the
              following general form:

              matrixName option ?arg arg ...?
                     Option and the args determine the exact behavior of the command.

       If matrixName is not specified a unique name will be generated by the package  itself.  If  a  source  is
       specified  the  new matrix will be initialized to it. For the operators =, :=, and as the argument source
       is interpreted as the name of another matrix object, and the assignment operator = will be executed.  For
       deserialize the source is a serialized matrix object and deserialize will be executed.

       In other words

              ::struct::matrix mymatrix = b

       is equivalent to

              ::struct::matrix mymatrix
              mymatrix = b

       and

              ::struct::matrix mymatrix deserialize $b

       is equivalent to

              ::struct::matrix mymatrix
              mymatrix deserialize $b

       The following commands are possible for matrix objects:

       matrixName = sourcematrix
              This  is  the assignment operator for matrix objects. It copies the matrix contained in the matrix
              object sourcematrix over the matrix data in matrixName. The old contents of matrixName are deleted
              by this operation.

              This operation is in effect equivalent to

              matrixName deserialize [sourcematrix serialize]

       matrixName --> destmatrix
              This  is the reverse assignment operator for matrix objects. It copies the matrix contained in the
              matrix object matrixName over the matrix data in the  object  destmatrix.   The  old  contents  of
              destmatrix are deleted by this operation.

              This operation is in effect equivalent to

              destmatrix deserialize [matrixName serialize]

       matrixName add column ?values?
              Extends  the  matrix by one column and then acts like set column (see below) on this new column if
              there were values supplied. Without values the new cells will be set to the empty string. The  new
              column is appended immediately behind the last existing column.

       matrixName add row ?values?
              Extends the matrix by one row and then acts like set row (see below) on this new row if there were
              values supplied. Without values the new cells will be set to the empty  string.  The  new  row  is
              appended immediately behind the last existing row.

       matrixName add columns n
              Extends  the  matrix  by n columns. The new cells will be set to the empty string. The new columns
              are appended immediately behind the last existing column. A value of n equal to or smaller than  0
              is not allowed.

       matrixName add rows n
              Extends  the  matrix  by  n  rows. The new cells will be set to the empty string. The new rows are
              appended immediately behind the last existing row. A value of n equal to or smaller than 0 is  not
              allowed.

       matrixName cells
              Returns  the  number  of  cells  currently  managed by the matrix. This is the product of rows and
              columns.

       matrixName cellsize column row
              Returns the length of the string representation of the value currently contained in the  addressed
              cell.

       matrixName columns
              Returns the number of columns currently managed by the matrix.

       matrixName columnwidth column
              Returns  the  length of the longest string representation of all the values currently contained in
              the cells of the addressed column if these are  all  spanning  only  one  line.  For  cell  values
              spanning multiple lines the length of their longest line goes into the computation.

       matrixName delete column column
              Deletes  the specified column from the matrix and shifts all columns with higher indices one index
              down.

       matrixName delete columns n
              Deletes n columns from the right of the matrix. The value of n has to satisfy the constraint "0  <
              n < [matrixName columns]"

       matrixName delete row row
              Deletes the specified row from the matrix and shifts all row with higher indices one index down.

       matrixName delete rows n
              Deletes  n rows from the bottom of the matrix. The value of n has to satisfy the constraint "0 < n
              < [matrixName rows]"

       matrixName deserialize serialization
              This is the complement to serialize. It  replaces  matrix  data  in  matrixName  with  the  matrix
              described  by  the  serialization  value.  The  old  contents  of  matrixName  are deleted by this
              operation.

       matrixName destroy
              Destroys the matrix, including its storage space and associated command.

       matrixName format 2string ?report?
              Formats the matrix using the specified report object and returns the string containing the  result
              of this operation. The report has to support the printmatrix method. If no report is specified the
              system will use an internal report definition to format the matrix.

       matrixName format 2chan ??report? channel?
              Formats the matrix using the specified report object and writes the string containing  the  result
              of  this operation into the channel. The report has to support the printmatrix2channel method.  If
              no report is specified the system will use an internal report definition to format the matrix.  If
              no channel is specified the system will use stdout.

       matrixName get cell column row
              Returns the value currently contained in the cell identified by row and column index.

       matrixName get column column
              Returns  a  list  containing  the values from all cells in the column identified by the index. The
              contents of the cell in row 0 are stored as the first element of this list.

       matrixName get rect column_tl row_tl column_br row_br
              Returns a list of lists of cell values. The values stored in the result come from  the  sub-matrix
              whose  top-left  and  bottom-right  cells are specified by column_tl, row_tl and column_br, row_br
              resp. Note that the following equations have to be true: "column_tl <= column_br" and  "row_tl  <=
              row_br".  The result is organized as follows: The outer list is the list of rows, its elements are
              lists representing a single row. The row with the smallest index is the first element of the outer
              list. The elements of the row lists represent the selected cell values. The cell with the smallest
              index is the first element in each row list.

       matrixName get row row
              Returns a list containing the values from all cells in  the  row  identified  by  the  index.  The
              contents of the cell in column 0 are stored as the first element of this list.

       matrixName insert column column ?values?
              Extends  the  matrix by one column and then acts like set column (see below) on this new column if
              there were values supplied. Without values the new cells will be set to the empty string. The  new
              column  is  inserted just before the column specified by the given index. This means, if column is
              less than or equal to zero, then the new column is inserted at the beginning of the matrix, before
              the  first column. If column has the value end, or if it is greater than or equal to the number of
              columns in the matrix, then the new column is appended to the matrix, behind the last column.  The
              old column at the chosen index and all columns with higher indices are shifted one index upward.

       matrixName insert row row ?values?
              Extends the matrix by one row and then acts like set row (see below) on this new row if there were
              values supplied. Without values the new cells will be set to the empty  string.  The  new  row  is
              inserted  just  before  the  row  specified by the given index. This means, if row is less than or
              equal to zero, then the new row is inserted at the beginning of the matrix, before the first  row.
              If  row  has the value end, or if it is greater than or equal to the number of rows in the matrix,
              then the new row is appended to the matrix, behind the last row. The old row at that index and all
              rows with higher indices are shifted one index upward.

       matrixName link ?-transpose? arrayvar
              Links the matrix to the specified array variable. This means that the contents of all cells in the
              matrix is stored in the array too, with all changes  to  the  matrix  propagated  there  too.  The
              contents  of  the cell (column,row) is stored in the array using the key column,row. If the option
              -transpose is specified the key row,column will be used instead. It is possible to link the matrix
              to  more  than  one  array.  Note  that  the  link is bidirectional, i.e. changes to the array are
              mirrored in the matrix too.

       matrixName links
              Returns a list containing the names of all array variables the matrix was linked to through a call
              to method link.

       matrixName rowheight row
              Returns the height of the specified row in lines. This is the highest number of lines spanned by a
              cell over all cells in the row.

       matrixName rows
              Returns the number of rows currently managed by the matrix.

       matrixName search ?-nocase? ?-exact|-glob|-regexp? all pattern
              Searches the whole matrix for cells matching the pattern and returns a list with all matches. Each
              item  in  the  aforementioned  list  is a list itself and contains the column and row index of the
              matching cell, in this order. The results are ordered by column first and row second,  both  times
              in  ascending  order.  This  means  that matches to the left and the top of the matrix come before
              matches to the right and down.

              The type of the pattern (string, glob, regular expression) is determined by the option  after  the
              search keyword. If no option is given it defaults to -exact.

              If the option -nocase is specified the search will be case-insensitive.

       matrixName search ?-nocase? ?-exact|-glob|-regexp? column column pattern
              Like search all, but the search is restricted to the specified column.

       matrixName search ?-nocase? ?-exact|-glob|-regexp? row row pattern
              Like search all, but the search is restricted to the specified row.

       matrixName search ?-nocase? ?-exact|-glob|-regexp? rect column_tl row_tl column_br row_br pattern
              Like search all, but the search is restricted to the specified rectangular area of the matrix.

       matrixName serialize ?column_tl row_tl column_br row_br?
              This method serializes the sub-matrix spanned up by the rectangle specification. In other words it
              returns a tcl value completely describing that matrix. If no  rectangle  is  specified  the  whole
              matrix  will  be  serialized.   This allows, for example, the transfer of matrix objects (or parts
              thereof) over arbitrary channels, persistence, etc.  This method is also the basis  for  both  the
              copy constructor and the assignment operator.

              The  result of this method has to be semantically identical over all implementations of the matrix
              interface. This is what will enable us to copy matrix data between  different  implementations  of
              the same interface.

              The result is a list containing exactly three items.

              The  first  two elements of the list specify the number of rows and columns of the matrix, in that
              order. The values integer numbers greater than or equal to zero.

              The last element of the list contains the values of the matrix cells we have  serialized,  in  the
              form  of  a value like it is returned by the get rect. However empty cells to the right and bottom
              of the matrix can be left out of that value as the size information in  the  serialization  allows
              the receiver the creation of a matrix with the proper size despite the missing values.
              # A possible serialization for the matrix structure
              #
              # | a b d g |
              # | c e     |
              # | f       |
              #
              # is
              #
              # 3 4 {{a b d g} {c e} {f}}

       matrixName set cell column row value
              Sets the value in the cell identified by row and column index to the data in the third argument.

       matrixName set column column values
              Sets  the  values in the cells identified by the column index to the elements of the list provided
              as the third argument. Each element of the list is assigned to one cell, with  the  first  element
              going  into the cell in row 0 and then upward. If there are less values in the list than there are
              rows the remaining rows are set to the empty string. If there are more values  in  the  list  than
              there are rows the superfluous elements are ignored. The matrix is not extended by this operation.

       matrixName set rect column row values
              Takes  a  list  of  lists of cell values and writes them into the submatrix whose top-left cell is
              specified by the two indices. If the sublists of the outerlist are not of equal length the shorter
              sublists  will be filled with empty strings to the length of the longest sublist. If the submatrix
              specified by the top-left cell and the number of rows and columns in the values extends beyond the
              matrix  we  are modifying the over-extending parts of the values are ignored, i.e. essentially cut
              off. This subcommand expects its input in the format as returned by getrect.

       matrixName set row row values
              Sets the values in the cells identified by the row index to the elements of the list  provided  as
              the third argument. Each element of the list is assigned to one cell, with the first element going
              into the cell in column 0 and then upward. If there are less values in the  list  than  there  are
              columns  the  remaining  columns are set to the empty string. If there are more values in the list
              than there are columns the superfluous elements are ignored. The matrix is not  extended  by  this
              operation.

       matrixName sort columns ?-increasing|-decreasing? row
              Sorts  the  columns  in  the matrix using the data in the specified row as the key to sort by. The
              options -increasing and -decreasing have the same meaning as for lsort.  If no option is specified
              -increasing is assumed.

       matrixName sort rows ?-increasing|-decreasing? column
              Sorts  the  rows  in  the matrix using the data in the specified column as the key to sort by. The
              options -increasing and -decreasing have the same meaning as for lsort.  If no option is specified
              -increasing is assumed.

       matrixName swap columns column_a column_b
              Swaps the contents of the two specified columns.

       matrixName swap rows row_a row_b
              Swaps the contents of the two specified rows.

       matrixName transpose
              Transposes the contents of the matrix, i.e. swaps rows for columns and vice versa.

       matrixName unlink arrayvar
              Removes the link between the matrix and the specified arrayvariable, if there is one.

EXAMPLES

       The  examples  below assume a 5x5 matrix M with the first row containing the values 1 to 5, with 1 in the
       top-left cell. Each other row contains the contents of the row above it,  rotated  by  one  cell  to  the
       right.

              % M getrect 0 0 4 4
              {{1 2 3 4 5} {5 1 2 3 4} {4 5 1 2 3} {3 4 5 1 2} {2 3 4 5 1}}

              % M setrect 1 1 {{0 0 0} {0 0 0} {0 0 0}}
              % M getrect 0 0 4 4
              {{1 2 3 4 5} {5 0 0 0 4} {4 0 0 0 3} {3 0 0 0 2} {2 3 4 5 1}}

       Assuming  that  the  style  definitions  in the example section of the manpage for the package report are
       loaded into the interpreter now an example which formats a matrix into a tabular report. The code filling
       the matrix with data is not shown.  contains useful data.

              % ::struct::matrix m
              % # ... fill m with data, assume 5 columns
              % ::report::report r 5 style captionedtable 1
              % m format 2string r
              +---+-------------------+-------+-------+--------+
              |000|VERSIONS:          |2:8.4a3|1:8.4a3|1:8.4a3%|
              +---+-------------------+-------+-------+--------+
              |001|CATCH return ok    |7      |13     |53.85   |
              |002|CATCH return error |68     |91     |74.73   |
              |003|CATCH no catch used|7      |14     |50.00   |
              |004|IF if true numeric |12     |33     |36.36   |
              |005|IF elseif          |15     |47     |31.91   |
              |   |true numeric       |       |       |        |
              +---+-------------------+-------+-------+--------+
              %
              % # alternate way of doing the above
              % r printmatrix m

BUGS, IDEAS, FEEDBACK

       This  document,  and  the package it describes, will undoubtedly contain bugs and other problems.  Please
       report   such    in    the    category    struct    ::    matrix    of    the    Tcllib    SF    Trackers
       [http://sourceforge.net/tracker/?group_id=12883].   Please also report any ideas for enhancements you may
       have for either package and/or documentation.

KEYWORDS

       matrix

CATEGORY

       Data structures

COPYRIGHT

       Copyright (c) 2002 Andreas Kupries <andreas_kupries@users.sourceforge.net>