# Created by Octave 3.2.4, Thu May 13 21:59:34 2010 UTC <root@voltaire>
# name: cache
# type: cell
# rows: 3
# columns: 20
# name: <cell-element>
# type: string
# elements: 1
# length: 11
append_save
# name: <cell-element>
# type: string
# elements: 1
# length: 17
 Input checking:

# name: <cell-element>
# type: string
# elements: 1
# length: 17
 Input checking:


# name: <cell-element>
# type: string
# elements: 1
# length: 15
calccelladdress
# name: <cell-element>
# type: string
# elements: 1
# length: 239
 calccelladdress (R, C) - compute spreadsheet style cell address from
 row & column index (both 1-based).
 
 Max column index currently set to 18278 (max ODS: 1024, OOXML: 16384).
 Row limits for ODF and OOXML are 65536 and 1048576, resp.

# name: <cell-element>
# type: string
# elements: 1
# length: 80
 calccelladdress (R, C) - compute spreadsheet style cell address from
 row & col

# name: <cell-element>
# type: string
# elements: 1
# length: 6
fexist
# name: <cell-element>
# type: string
# elements: 1
# length: 680
 -- Function File: ex = fexist (file, tspec, aspec)
     Checks whether a file exists.  FILE is the queried file path.
     TSPEC is a combination of letters f,d,p,S, corresponding to file
     types:
        * f: regular file

        * d: directory

        * p: named pipe (FIFO special file)

        * S: socket

     The query is true if the actual file type matches any of the
     specified options.

     ASPEC is a combination of letters r,w,x, corresponding to queried
     access privileges to the file. The query is true if the current
     user has all the spefied types of access, either through "user",
     "group" or "other" specs.

     See also: stat, lstat



# name: <cell-element>
# type: string
# elements: 1
# length: 29
Checks whether a file exists.

# name: <cell-element>
# type: string
# elements: 1
# length: 12
getusedrange
# name: <cell-element>
# type: string
# elements: 1
# length: 50
 getusedrange - get used range from a spreadsheet

# name: <cell-element>
# type: string
# elements: 1
# length: 50
 getusedrange - get used range from a spreadsheet


# name: <cell-element>
# type: string
# elements: 1
# length: 7
oct2ods
# name: <cell-element>
# type: string
# elements: 1
# length: 2289
 -- Function File: [ ODS, RSTATUS ] = oct2ods (ARR, ODS)
 -- Function File: [ ODS, RSTATUS ] = oct2ods (ARR, ODS, WSH)
 -- Function File: [ ODS, RSTATUS ] = oct2ods (ARR, ODS, WSH, RANGE)
     Transfer data to an OpenOffice_org Calc spreadsheet previously
     opened by odsopen().

     Data in 1D/2D array ARR are transferred into a cell range RANGE in
     sheet WSH. ODS must have been made earlier by odsopen(). Return
     argument ODS should be the same as supplied argument ODS and is
     updated by oct2ods. A subsequent call to odsclose is needed to
     write the updated spreadsheet to disk (and -if needed- close the
     Java invocation holding the file pointer).

     ARR can be any array type save complex. Mixed numeric/text arrays
     can only be cell arrays.

     ODS must be a valid pointer struct created earlier by odsopen.

     WSH can be a number or string.  In case of a yet non-existing Calc
     file, the first sheet will be used & named according to WSH.  In
     case of existing files, some checks are made for existing sheet
     names or numbers.  When new sheets are to be added to the Calc
     file, they are inserted to the right of all existing sheets. The
     pointer to the "active" sheet (shown when Calc opens the file)
     remains untouched.

     If omitted, RANGE is initially supposed to be 'A1:AMJ65536'.  The
     actual range to be used is determined by the size of ARR.  Be
     aware that data array sizes > 2.10^5 elements may exhaust the java
     shared memory space for the default java memory settings.  For
     larger arrays appropriate memory settings are needed in the file
     java.opts; then the maximum array size for the java-based
     spreadsheet options is about 5-6.10^5 elements.

     Data are added to the sheet, ignoring other data already present;
     existing data in the range to be used will be overwritten.

     If RANGE contains merged cells, also the elements of ARR not
     corresponding to the top or left Calc cells of those merged cells
     will be written, however they won't be shown until in Calc the
     merge is undone.

     Examples:

            [ods, status] = ods2oct ('arr', ods, 'Newsheet1', 'AA31:GH165');

     See also: ods2oct, odsopen, odsclose, odsread, odswrite, odsfinfo



# name: <cell-element>
# type: string
# elements: 1
# length: 80
Transfer data to an OpenOffice_org Calc spreadsheet previously opened
by odsopen

# name: <cell-element>
# type: string
# elements: 1
# length: 7
oct2xls
# name: <cell-element>
# type: string
# elements: 1
# length: 2575
 -- Function File: [ XLS, RSTATUS ] = oct2xls (ARR, XLS)
 -- Function File: [ XLS, RSTATUS ] = oct2xls (ARR, XLS, WSH)
 -- Function File: [ XLS, RSTATUS ] = oct2xls (ARR, XLS, WSH, TOPLEFT)
     Add data in 1D/2D CELL array ARR into a range with upper left cell
     equal to TOPLEFT in worksheet WSH in an Excel spreadsheet file
     pointed to in structure XLS.  Return argument XLS equals supplied
     argument XLS and is updated by oct2xls.

     A subsequent call to xlsclose is needed to write the updated
     spreadsheet to disk (and -if needed- close the Excel or Java
     invocation).

     ARR can be any array type save complex. Mixed numeric/text arrays
     can only be cell arrays.

     XLS must be a valid pointer struct created earlier by xlsopen.

     WSH can be a number or string (max. 31 chars).  In case of a yet
     non-existing Excel file, the first worksheet will be used & named
     according to WSH - the extra worksheets that Excel creates by
     default are deleted.  In case of existing files, some checks are
     made for existing worksheet names or numbers, or whether WSH
     refers to an existing sheet with a type other than worksheet
     (e.g., chart).  When new worksheets are to be added to the Excel
     file, they are inserted to the right of all existing worksheets.
     The pointer to the "active" sheet (shown when Excel opens the
     file) remains untouched.

     If omitted, TOPLEFT is supposed to be 'A1'. The actual range to be
     used is determined by the size of ARR.

     Data are added to the worksheet, ignoring other data already
     present; existing data in the range to be used will be overwritten.

     If RANGE contains merged cells, only the elements of ARR
     corresponding to the top or left Excel cells of those merged cells
     will be written, other array cells corresponding to that cell will
     be ignored.

     Beware that -if invoked- Excel invocations may be left running
     silently in case of COM errors. Invoke xlsclose with proper
     pointer struct to close them.  When using java, note that data
     array sizes > 2.10^5 elements may exhaust the java shared memory
     space for the default java memory settings.  For larger arrays
     appropriate memory settings are needed in the file java.opts; then
     the maximum array size for the java-based spreadsheet options is
     about 5-6.10^5 elements.

     Examples:

            [xlso, status] = xls2oct ('arr', xlsi, 'Third_sheet', 'AA31');

     See also: xls2oct, xlsopen, xlsclose, xlsread, xlswrite, xlsfinfo



# name: <cell-element>
# type: string
# elements: 1
# length: 80
Add data in 1D/2D CELL array ARR into a range with upper left cell
equal to TOPL

# name: <cell-element>
# type: string
# elements: 1
# length: 7
ods2oct
# name: <cell-element>
# type: string
# elements: 1
# length: 2952
 -- Function File: [ RAWARR, ODS, RSTATUS ] = ods2oct (ODS)
 -- Function File: [ RAWARR, ODS, RSTATUS ] = ods2oct (ODS, WSH)
 -- Function File: [ RAWARR, ODS, RSTATUS ] = ods2oct (ODS, WSH, RANGE)
     Read data contained within range RANGE from worksheet WSH in an
     OpenOffice.org spreadsheet file pointed to in struct ODS.

     ods2oct is a mere wrapper for interface-dependent scripts (e.g.,
     ods2jotk2oct and ods2jod2oct) that do the actual reading.

     WSH is either numerical or text, in the latter case it is
     case-sensitive and it may be max. 31 characters long.  Note that
     in case of a numerical WSH this number refers to the position in
     the worksheet stack, counted from the left in a Calc window. The
     default is numerical 1, i.e. the leftmost worksheet in the ODS
     file.

     RANGE is expected to be a regular spreadsheet range format, or ""
     (empty string, indicating all data in a worksheet).

     If only the first argument is specified, ods2oct will try to read
     all contents from the first = leftmost (or the only) worksheet (as
     if a range of '' (empty string) was specified).

     If only two arguments are specified, ods2oct assumes the second
     argument to be WSH. In that case ods2oct will try to read all data
     contained in that worksheet.

     Return argument RAWARR contains the raw spreadsheet cell data.
     Optional return argument ODS contains the pointer struct. Field
     ODS.limits contains the outermost column and row numbers of the
     actually read cell range.  RSTATUS will be set to 1 if the
     requested data have been read successfully, 0 otherwise.  Use
     parsecell() to separate numeric and text values from RAWARR.

     ODS is supposed to have been created earlier by odsopen in the
     same octave session. It is only referred to, not changed.

     Erroneous data and empty cells turn up empty in RAWARR.  Date/time
     values in OpenOffice.org are returned as numerical values with
     base 1-1-0000 (same as octave). But beware that Excel spreadsheets
     rewritten by OpenOffice.org into .ods format may have numerical
     date cells with base 01-01-1900 (same as MS-Excel).

     Be aware that ods2oct trims RAWARR from empty outer rows & columns,
     so any returned cell array may turn out to be smaller than
     requested in RANGE.

     When reading from merged cells, all array elements NOT
     corresponding to the leftmost or upper OpenOffice.org cell will be
     treated as if the "corresponding" cells are empty.

     Examples:

            A = ods2oct (ods1, '2nd_sheet', 'C3:ABS40000');
            (which returns the numeric contents in range C3:ABS40000 in worksheet
            '2nd_sheet' from a spreadsheet file pointed to in pointer struct ods1,
            into numeric array A)

            [An, ods2, status] = ods2oct (ods2, 'Third_sheet');

     See also: odsopen, odsclose, parsecell, odsread, odsfinfo



# name: <cell-element>
# type: string
# elements: 1
# length: 75
Read data contained within range RANGE from worksheet WSH in an
OpenOffice.

# name: <cell-element>
# type: string
# elements: 1
# length: 8
odsclose
# name: <cell-element>
# type: string
# elements: 1
# length: 100
 odsclose - close an ods (OpenOffice_org) spreadsheet file and
 write it to disk if it was changed.

# name: <cell-element>
# type: string
# elements: 1
# length: 80
 odsclose - close an ods (OpenOffice_org) spreadsheet file and
 write it to disk

# name: <cell-element>
# type: string
# elements: 1
# length: 8
odsfinfo
# name: <cell-element>
# type: string
# elements: 1
# length: 1634
 -- Function File: [ FILETYPE] = odsfinfo (FILENAME [, REQINTF])
 -- Function File: [ FILETYPE, SH_NAMES] = odsfinfo (FILENAME [,
          REQINTF])
     Query an OpenOffice_org spreadsheet file FILENAME (with ods
     suffix) for some info about its contents.

     If FILENAME is a recognizable OpenOffice.org spreadsheet file,
     FILETYPE returns the string "OpenOffice.org Calc spreadsheet", or
     '' (empty string) otherwise.

     If FILENAME is a recognizable OpenOffice.org Calc spreadsheet
     file, optional argument SH_NAMES contains a list (cell array) of
     sheet names contained in FILENAME, in the order (from left to
     right) in which they occur in the sheet stack.

     If you omit return arguments FILETYPE and SH_NAMES altogether,
     odsfinfo returns the sheet names + (in case of the ODF toolkit
     interface) a qualified guess for the actual occupied data range to
     the screen.

     odsfinfo execution can take its time for large spreadsheets as the
     entire spreadsheet has to be parsed to get the sheet names, let
     alone exploring used data ranges.

     By specifying a value of 'jod' or 'otk' for REQINTF the automatic
     selection of the java interface is bypassed and the specified
     interface will be used (if at all present).

     Examples:

            exist = odsfinfo ('test4.ods');
            (Just checks if file test4.ods is a readable Calc file)

            [exist, names] = odsfinfo ('test4.ods');
            (Checks if file test4.ods is a readable Calc file and return a
             list of sheet names)

     See also: odsread, odsopen, ods2oct, odsclose



# name: <cell-element>
# type: string
# elements: 1
# length: 80
Query an OpenOffice_org spreadsheet file FILENAME (with ods suffix) for
some inf

# name: <cell-element>
# type: string
# elements: 1
# length: 7
odsopen
# name: <cell-element>
# type: string
# elements: 1
# length: 1666
 -- Function File: ODS = odsopen (FILENAME)
 -- Function File: ODS = odsopen (FILENAME, READWRITE)
 -- Function File: ODS = odsopen (FILENAME, READWRITE, REQINTF)
     Get a pointer to an OpenOffice_org spreadsheet in the form of
     return argument ODS.

     Calling odsopen without specifying a return argument is fairly
     useless!

     To make this function work at all, you need the Java package >
     1.2.6 plus ODFtoolkit version 0.7.5 & xercesImpl, and/or
     jOpenDocument installed on your computer + proper javaclasspath
     set. These interfaces are referred to as OTK and JOD, resp., and
     are preferred in that order by default (depending on their
     presence).

     FILENAME must be a valid .ods OpenOffice.org file name. If FILENAME
     does not contain any directory path, the file is saved in the
     current directory.

     READWRITE must be set to 1 if writing to spreadsheet is desired
     immediately after calling odsopen(). It merely serves proper
     handling of file errors (e.g., "file not found" or "new file
     created").

     Optional input argument REQINTF can be used to override the ODS
     interface automatically selected by odsopen. Currently implemented
     interfaces are 'OTK' (Java / ODFtoolkit) or 'JOD' (Java /
     jOpenDocument).

     Examples:

            ods = odsopen ('test1.ods');
            (get a pointer for reading from spreadsheet test1.ods)

            ods = odsopen ('test2.ods', [], 'JOD');
            (as above, indicate test2.ods will be read from; in this case using
             the jOpenDocument interface is requested)

     See also: odsclose, odsread, ods2oct, odsfinfo



# name: <cell-element>
# type: string
# elements: 1
# length: 80
Get a pointer to an OpenOffice_org spreadsheet in the form of return
argument OD

# name: <cell-element>
# type: string
# elements: 1
# length: 7
odsread
# name: <cell-element>
# type: string
# elements: 1
# length: 4031
 -- Function File: [ NUMARR, TXTARR, RAWARR, LIMITS] = odsread
          (FILENAME)
 -- Function File: [ NUMARR, TXTARR, RAWARR, LIMITS] = odsread
          (FILENAME, WSH)
 -- Function File: [ NUMARR, TXTARR, RAWARR, LIMITS] = odsread
          (FILENAME, WSH, RANGE)
 -- Function File: [ NUMARR, TXTARR, RAWARR, LIMITS] = odsread
          (FILENAME, WSH, RANGE, REQINTF)
     Read data contained in range RANGE from worksheet WSH in
     OpenOffice_org Calc spreadsheet file FILENAME.

     You need the octave-forge java package (> 1.2.6) and one or both of
     jopendocument.jar or preferrably: (odfdom.jar 0.7.5 &
     xercesImpl.jar) in your javaclasspath.

     Return argument NUMARR contains the numeric data, optional return
     arguments TXTARR and RAWARR contain text strings and the raw
     spreadsheet cell data, respectively, and LIMITS is a struct
     containing the data origins of the various returned arrays.

     If FILENAME does not contain any directory, the file is assumed to
     be in the current directory.

     WSH is either numerical or text, in the latter case it is
     case-sensitive and it should conformtoOpenOffice.org Calc sheet
     name requirements.  Note that in case of a numerical WSH this
     number refers to the position in the worksheet stack, counted from
     the left in a Calc window. The default is numerical 1, i.e. the
     leftmost worksheet in the Calc file.

     RANGE is expected to be a regular spreadsheet range format, or ""
     (empty string, indicating all data in a worksheet).

     If only the first argument is specified, odsread will try to read
     all contents from the first = leftmost (or the only) worksheet (as
     if a range of '' (empty string) was specified).

     If only two arguments are specified, odsread assumes the second
     argument to be WSH and to refer to a worksheet. In that case
     odsread tries to read all data contained in that worksheet.

     The optional last argument REQINTF can be used to override the
     automatic selection by odsread of one interface out of the
     supported ones: Java/ODFtoolkit ('OTK') or Java/jOpenDocument
     ('JOD').

     Erroneous data and empty cells are set to NaN in NUMARR and turn
     up empty in TXTARR and RAWARR. Date/time values in date/time
     formatted cells are returned as numerical values in OBJ with base
     1-1-000. Note that OpenOfice.org and MS-Excel have different date
     base values (1/1/0000 & 1/1/1900, resp.) and internal
     representation so MS-Excel spreadsheets rewritten into .ods format
     by OpenOffice.org Calc may have different date base values.

     NUMARR and TXTARR are trimmed from empty outer rows and columns,
     so any returned array may turn out to be smaller than requested in
     RANGE.

     When reading from merged cells, all array elements NOT
     corresponding to the leftmost or upper Calc cell will be treated
     as if the "corresponding" Calc cells are empty.

     odsread is just a wrapper for a collection of scripts that find out
     the interface to be used and do the actual reading. For each call
     to odsread the interface must be started and the Calc file read
     into memory. When reading multiple ranges (in optionally multiple
     worksheets) a significant speed boost can be obtained by invoking
     those scripts directly (odsopen / ods2oct [/ parsecell] / ... /
     odsclose).

     Examples:

            A = odsread ('test4.ods', '2nd_sheet', 'C3:AB40');
            (which returns the numeric contents in range C3:AB40 in worksheet
            '2nd_sheet' from file test4.ods into numeric array A)

            [An, Tn, Ra, limits] = odsread ('Sales2009.ods', 'Third_sheet');
            (which returns all data in worksheet 'Third_sheet' in file test4.ods
            into array An, the text data into array Tn, the raw cell data into
            cell array Ra and the ranges from where the actual data came in limits)

     See also: odsopen, ods2oct, odsclose, odsfinfo, parsecell



# name: <cell-element>
# type: string
# elements: 1
# length: 80
Read data contained in range RANGE from worksheet WSH in OpenOffice_org
Calc spr

# name: <cell-element>
# type: string
# elements: 1
# length: 8
odswrite
# name: <cell-element>
# type: string
# elements: 1
# length: 2984
 -- Function File: RSTATUS = odswrite (FILENAME, ARR)
 -- Function File: RSTATUS = odswrite (FILENAME, ARR, WSH)
 -- Function File: RSTATUS = odswrite (FILENAME, ARR, WSH, RANGE)
 -- Function File: RSTATUS = odswrite (FILENAME, ARR, WSH, RANGE,
          REQINTF)
     Add data in 1D/2D array ARR to sheet WSH in OpenOffice_org Calc
     spreadsheet file FILENAME in range RANGE.

     RSTATUS returns 1 if write succeeded, 0 otherwise.

     FILENAME must be a valid .ods OpenOffice.org file name. If
     FILENAME does not contain any directory path, the file is saved in
     the current directory.

     ARR can be any array type save complex. Mixed numeric/text arrays
     can only be cell arrays.

     WSH can be a number or string (max. 31 chars).  In case of a not
     yet existing OpenOffice.org spreadsheet, the first sheet will be
     used & named according to WSH - no extra empty sheets are created.
     In case of existing files, some checks are made for existing sheet
     names or numbers, or whether WSH refers to an existing sheet with
     a type other than sheet (e.g., chart).  When new sheets are to be
     added to the spreadsheet file, they are inserted to the right of
     all existing sheets. The pointer to the "active" sheet (shown when
     OpenOffice.org Calc opens the file) remains untouched.

     RANGE is expected to be a regular spreadsheet range.  Data is
     added to the sheet; existing data in the requested range will be
     overwritten.  Array ARR will be clipped at the right and/or bottom
     if its size is bigger than can be accommodated in RANGE.  If ARR
     is smaller than the RANGE allows, it is placed in the top left of
     RANGE.

     If RANGE contains merged cells, only the elements of ARR
     corresponding to the top or left Calc cells of those merged cells
     will be written, other array cells corresponding to that cell will
     be ignored.

     The optional last argument REQINTF can be used to override the
     automatic selection by odswrite of one interface out of the
     supported ones: Java/ODFtooolkit ('OTK'), or Java/jOpenDocument
     ('JOD').

     odswrite is a mere wrapper for various scripts which find out what
     ODS interface to use (ODF toolkit or jOpenDocument) plus code to
     mimic the other brand's syntax. For each call to odswrite such an
     interface must be started and possibly an ODS file loaded. When
     writing to multiple ranges and/or worksheets in the same ODS file,
     a speed bonus can be obtained by invoking those scripts (odsopen /
     octods / .... / odsclose) directly.

     Example:

            status = odswrite ('test4.ods', 'arr', 'Eight_sheet', 'C3:AB40');
            (which adds the contents of array arr (any type) to range C3:AB40
            in sheet 'Eight_sheet' in file test4.ods and returns a logical
            True (= numerical 1) in status if al went well)

     See also: odsread, oct2ods, ods2oct, odsopen, odsclose, odsfinfo



# name: <cell-element>
# type: string
# elements: 1
# length: 80
Add data in 1D/2D array ARR to sheet WSH in OpenOffice_org Calc
spreadsheet file

# name: <cell-element>
# type: string
# elements: 1
# length: 14
parse_sp_range
# name: <cell-element>
# type: string
# elements: 1
# length: 199
 Parse a string representing a range of cells for a spreadsheet
 into nr of rows and nr of columns and also extract top left
 cell address + top row + left column. Some error checks are implemented.

# name: <cell-element>
# type: string
# elements: 1
# length: 80
 Parse a string representing a range of cells for a spreadsheet
 into nr of rows

# name: <cell-element>
# type: string
# elements: 1
# length: 9
parsecell
# name: <cell-element>
# type: string
# elements: 1
# length: 1359
 -- Function File: [ NUMARR, TXTARR, LIM ] = parsecell (RAWARR)
 -- Function File: [ NUMARR, TXTARR, LIM ] = parsecell (RAWARR, LIMITS)
     Divide a heterogeneous 2D cell array into a 2D numeric array and a
     2D cell array containing only strings. Both returned arrays are
     trimmed from empty outer rows and columns.  This function is
     particularly useful for parsing cell arrays returned by functions
     reading spreadsheets (e.g., xlsread, odsread).

     Optional return argument LIM contains two field with the outer
     column and row numbers of NUMARR and TXTARR in the original array
     RAWARR.  If optional input argument LIMITS contained the
     spreadsheet data limits returned in the spreadsheet file pointer
     struct (field xls.limits or ods.limits), optional return argument
     LIM contains the real spreadsheet row & column numbers enclosing
     the origins of the numerical and text data returned in NUMARR and
     TXTARR.

     Examples:

            [An, Tn] = parsecell (Rn);
            (which returns the numeric contents of Rn into array An and the
             text data into array Tn)

            [An, Tn, lims] = parsecell (Rn, xls.limits);
            (which returns the numeric contents of Rn into array An and the
             text data into array Tn.)

     See also: xlsread, odsread, xls2oct, ods2oct



# name: <cell-element>
# type: string
# elements: 1
# length: 80
Divide a heterogeneous 2D cell array into a 2D numeric array and a 2D
cell array

# name: <cell-element>
# type: string
# elements: 1
# length: 7
xls2oct
# name: <cell-element>
# type: string
# elements: 1
# length: 3161
 -- Function File: [ RAWARR, XLS, RSTATUS ] = xls2oct (XLS)
 -- Function File: [ RAWARR, XLS, RSTATUS ] = xls2oct (XLS, WSH)
 -- Function File: [ RAWARR, XLS, RSTATUS ] = xls2oct (XLS, WSH, RANGE)
     Read data contained within range RANGE from worksheet WSH in an
     Excel spreadsheet file pointed to in struct XLS.

     WSH is either numerical or text, in the latter case it is
     case-sensitive and it may be max. 31 characters long.  Note that
     in case of a numerical WSH this number refers to the position in
     the worksheet stack, counted from the left in an Excel window. The
     default is numerical 1, i.e. the leftmost worksheet in the Excel
     file.

     RANGE is expected to be a regular spreadsheet range format, or ""
     (empty string, indicating all data in a worksheet).

     If only the first argument is specified, xls2oct will try to read
     all contents from the first = leftmost (or the only) worksheet (as
     if a range of '' (empty string) was specified).

     If only two arguments are specified, xls2oct assumes the second
     argument to be WSH. In that case xls2oct will try to read all data
     contained in that worksheet.

     Return argument RAWARR contains the raw spreadsheet cell data.
     Optional return argument XLS contains the pointer struct, RSTATUS
     will be set to 1 if the requested data have been read
     successfully, 0 otherwise.  Use parsecell() to separate numeric
     and text values from RAWARR.

     XLS is supposed to have been created earlier by xlsopen in the
     same octave session. It is only referred to, not changed.

     If one of the Java interfaces is used, field XLS.limits contains
     the outermost column and row numbers of the actually read cell
     range.  This doesn't work with native Excel / COM.

     Erroneous data and empty cells turn up empty in RAWARR.  Date/time
     values in Excel are returned as numerical values.  Note that Excel
     and Octave have different date base values (1/1/1900 & 1/1/0000,
     resp.)  Be aware that Excel trims RAWARR from empty outer rows &
     columns, so any returned cell array may turn out to be smaller
     than requested in RANGE.  When using COM or POI interface,
     formulas in cells are evaluated; if that fails cached values are
     retrieved. Those may be outdated depending on Excel's "Automatic
     calculation" settings when the spreadsheet was saved.

     When reading from merged cells, all array elements NOT
     corresponding to the leftmost or upper Excel cell will be treated
     as if the "corresponding" Excel cells are empty.

     Beware: when the COM interface is used, hidden Excel invocations
     may be kept running silently in case of COM errors.

     Examples:

            A = xls2oct (xls1, '2nd_sheet', 'C3:AB40');
            (which returns the numeric contents in range C3:AB40 in worksheet
            '2nd_sheet' from a spreadsheet file pointed to in pointer struct xls1,
            into numeric array A)

            [An, xls2, status] = xls2oct (xls2, 'Third_sheet');

     See also: oct2xls, xlsopen, xlsclose, parsecell, xlsread,
     xlsfinfo, xlswrite



# name: <cell-element>
# type: string
# elements: 1
# length: 80
Read data contained within range RANGE from worksheet WSH in an Excel
spreadshee

# name: <cell-element>
# type: string
# elements: 1
# length: 8
xlsclose
# name: <cell-element>
# type: string
# elements: 1
# length: 878
 -- Function File: [ XLS] = xlsclose (XLS)
     Close the Excel spreadsheet pointed to in struct XLS, if needed
     write the file to disk. An empty pointer struct will be returned.
     xlsclose will determine if the file must be written to disk based
     on information contained in XLS.

     You need MS-Excel (95 - 2003), and/or the Java package > 1.2.6
     plus Apache POI > 3.5 installed on your computer + proper
     javaclasspath set, to make this function work at all.

     XLS must be a valid pointer struct made by xlsopen() in the same
     octave session.

     Beware: Excel invocations may be left running invisibly in case of
     COM errors.

     Examples:

            xls1 = xlsclose (xls1);
            (Close spreadsheet file pointed to in pointer struct xls1; xls1 is reset)

     See also: xlsclose, xlsread, xlswrite, xls2oct, oct2xls, xlsfinfo



# name: <cell-element>
# type: string
# elements: 1
# length: 80
Close the Excel spreadsheet pointed to in struct XLS, if needed write
the file t

# name: <cell-element>
# type: string
# elements: 1
# length: 8
xlsfinfo
# name: <cell-element>
# type: string
# elements: 1
# length: 1883
 -- Function File: [ FILETYPE] = xlsfinfo (FILENAME [, REQINTF])
 -- Function File: [ FILETYPE, SH_NAMES] = xlsfinfo (FILENAME [,
          REQINTF])
 -- Function File: [ FILETYPE, SH_NAMES, FFORMAT] = xlsfinfo (FILENAME
          [, REQINTF])
     Query Excel spreadsheet file FILENAME for some info about its
     contents.

     If FILENAME is a recognizable Excel spreadsheet file, FILETYPE
     returns the string "Microsoft Excel Spreadsheet", or '' (empty
     string) otherwise.

     If FILENAME is a recognizable Excel spreadsheet file, optional
     argument SH_NAMES contains a list (cell array) of sheet names (and
     in case Excel is installed: sheet types) contained in FILENAME, in
     the order (from left to right) in which they occur in the sheet
     stack.

     Optional return value FFORMAT currently returns '' (empty string)
     unless FILENAME is a readable Excel 95-2003 .xls file in which
     case FFORMAT is set to "xlWorkbookNormal".

     If no return arguments are specified the sheet names are echoed to
     the terminal screen; in case of java interfaces for each sheet the
     actual occupied data range is echoed as well.

     If multiple xls interfaces have been installed REQINTF can be
     specified. This can sometimes be handy to get an idea of used cell
     ranges in each worksheet (the COM/Excel interface can't supply
     this information).

     For use on OOXML spreadsheets one needs full POI support (see
     xlsopen) and 'poi' needs to be specified for REQINTF.

     Examples:

            exist = xlsfinfo ('test4.xls');
            (Just checks if file test4.xls is a readable Excel file)

            [exist, names] = xlsfinfo ('test4.xls');
            (Checks if file test4.xls is a readable Excel file and return a
             list of sheet names and -types)

     See also: oct2xls, xlsread, xls2oct, xlswrite



# name: <cell-element>
# type: string
# elements: 1
# length: 71
Query Excel spreadsheet file FILENAME for some info about its contents.

# name: <cell-element>
# type: string
# elements: 1
# length: 7
xlsopen
# name: <cell-element>
# type: string
# elements: 1
# length: 2117
 -- Function File: XLS = xlsopen (FILENAME)
 -- Function File: XLS = xlsopen (FILENAME, READWRITE)
 -- Function File: XLS = xlsopen (FILENAME, READWRITE, REQINTF)
     Get a pointer to an Excel spreadsheet in the form of return
     argument XLS.

     Calling xlsopen without specifying a return argument is fairly
     useless!

     To make this function work at all, you need MS-Excel (95 - 2003),
     and/or the Java package > 1.2.6 plus either Apache POI > 3.5 or
     JExcelAPI installed on your computer + proper javaclasspath set.
     These interfaces are referred to as COM, POI and JXL, resp., and
     are preferred in that order by default (depending on their
     presence).  For OOXML support, in addition to Apache POI support
     you also need the following jars in your javaclasspath:
     poi-ooxml-schemas-3.5.jar, xbean.jar and dom4j-1.6.1.jar (or later
     versions).

     FILENAME should be a valid .xls or xlsx Excel file name; but if
     you use the COM interface you can specify any extension that your
     installed Excel version can read AND write. If FILENAME does not
     contain any directory path, the file is saved in the current
     directory.

     If READWRITE is set to 0 (default value) or omitted, the Excel file
     is opened for reading. If READWRITE is set to True or 1, an Excel
     file is opened (or created) for reading & writing.

     Optional input argument REQINTF can be used to override the Excel
     interface automatically selected by xlsopen. Currently implemented
     interfaces are 'COM' (Excel / COM), 'POI' (Java / Apache POI) or
     'JXL' (Java / JExcelAPI).

     Beware: Excel invocations may be left running invisibly in case of
     COM errors.

     Examples:

            xls = xlsopen ('test1.xls');
            (get a pointer for reading from spreadsheet test1.xls)

            xls = xlsopen ('test2.xls', 1, 'POI');
            (as above, indicate test2.xls will be written to; in this case using Java
             and the Apache POI interface are requested)

     See also: xlsclose, xlsread, xlswrite, xls2oct, oct2xls, xlsfinfo



# name: <cell-element>
# type: string
# elements: 1
# length: 73
Get a pointer to an Excel spreadsheet in the form of return argument
XLS.

# name: <cell-element>
# type: string
# elements: 1
# length: 7
xlsread
# name: <cell-element>
# type: string
# elements: 1
# length: 4678
 -- Function File: [ NUMARR, TXTARR, RAWARR, LIMITS] = xlsread
          (FILENAME)
 -- Function File: [ NUMARR, TXTARR, RAWARR, LIMITS] = xlsread
          (FILENAME, WSH)
 -- Function File: [ NUMARR, TXTARR, RAWARR, LIMITS] = xlsread
          (FILENAME, RANGE)
 -- Function File: [ NUMARR, TXTARR, RAWARR, LIMITS] = xlsread
          (FILENAME, WSH, RANGE)
 -- Function File: [ NUMARR, TXTARR, RAWARR, LIMITS] = xlsread
          (FILENAME, WSH, RANGE, REQINTF)
     Read data contained in range RANGE from worksheet WSH in Excel
     spreadsheet file FILENAME.  Return argument NUMARR contains the
     numeric data, optional return arguments TXTARR and RAWARR contain
     text strings and the raw spreadsheet cell data, respectively.

     If neither Excel, Java/Apache POI or Java/JExcelAPI are installed,
     xlsread falls back to csv file reading.

     In case one of the Java interfaces (Apache POI / JExcelAPI) was
     invoked, LIMITS contains the outer column/row numbers of the
     spreadsheet range where NUMARR, TXTARR and RAWARR have come from
     (remember, xlsread trims outer rows and columns).

     If FILENAME does not contain any directory, the file is assumed to
     be in the current directory.

     RANGE is expected to be a regular spreadsheet range format, or ""
     (empty string, indicating all data in a worksheet).

     WSH is either numerical or text, in the latter case it is
     case-sensitive and it may be max. 31 characters long.  Note that
     in case of a numerical WSH this number refers to the position in
     the worksheet stack, counted from the left in an Excel window. The
     default is numerical 1, i.e. the leftmost worksheet in the Excel
     file.

     If only the first argument is specified, xlsread will try to read
     all contents from the first = leftmost (or the only) worksheet (as
     if a range of '' (empty string) was specified).

     If only two arguments are specified, xlsread assumes the second
     argument to be RANGE if it is a string argument and contains  a
     ":" or if it is '' (empty string) and in those cases assumes the
     data must be read from the first worksheet (not necessarily
     Sheet1! but the leftmost sheet).

     However, if only two arguments are specified and the second
     argument is numeric or a text string that does not contain a ":",
     it is assumed to be WSH and to refer to a worksheet. In that case
     xlsread tries to read all data contained in that worksheet.

     The optional last argument REQINTF can be used to override the
     automatic selection by xlsread of one interface out of the
     supported ones: COM/Excel, Java/Apache POI, or Java/JExcelAPI.
     For reading from OOXML files a value of 'poi' must be specified
     for REQINTF (see help for xlsopen).

     Erroneous data and empty cells are set to NaN in NUMARR and turn
     up empty in TXTARR and RAWARR. Date/time values in Excel are
     returned as numerical values in OBJ. Note that Excel and Octave
     have different date base values (1/1/1900 & 1/1/0000, resp.)
     NUMARR and TXTARR are trimmed from empty outer rows and columns.
     Be aware that Excel does the same for RAWARR, so any returned
     array may turn out to be smaller than requested in RANGE.

     When reading from merged cells, all array elements NOT
     corresponding to the leftmost or upper Excel cell will be treated
     as if the "corresponding" Excel cells are empty.

     xlsread is just a wrapper for a collection of scripts that find out
     the interface to be used (COM, Java/POI,Java/JXL) and do the actual
     reading. For each call to xlsread the interface must be started and
     the Excel file read into memory. When reading multiple ranges (in
     optionally multiple worksheets) a significant speed boost can be
     obtained by invoking those scripts directly (xlsopen /xls2oct / ...
     / xlsclose [/ parsecell]).

     Beware: when using the COM interface, hidden Excel invocations may
     be kept running silently if not closed explicitly.

     Examples:

            A = xlsread ('test4.xls', '2nd_sheet', 'C3:AB40');
            (which returns the numeric contents in range C3:AB40 in worksheet
            '2nd_sheet' from file test4.xls into numeric array A)

            [An, Tn, Ra, limits] = xlsread ('Sales2009.xls', 'Third_sheet');
            (which returns all data in worksheet 'Third_sheet' in file test4.xls
            into array An, the text data into array Tn, the raw cell data into
            cell array Ra and the ranges from where the actual data came in limits)

     See also: xlswrite, xlsopen, xls2oct, xlsclose, xlsfinfo, oct2xls



# name: <cell-element>
# type: string
# elements: 1
# length: 80
Read data contained in range RANGE from worksheet WSH in Excel
spreadsheet file 

# name: <cell-element>
# type: string
# elements: 1
# length: 8
xlswrite
# name: <cell-element>
# type: string
# elements: 1
# length: 3390
 -- Function File: RSTATUS = xlswrite (FILENAME, ARR)
 -- Function File: RSTATUS = xlswrite (FILENAME, ARR, WSH)
 -- Function File: RSTATUS = xlswrite (FILENAME, ARR, WSH, RANGE)
 -- Function File: RSTATUS = xlswrite (FILENAME, ARR, WSH, RANGE,
          REQINTF)
     Add data in 1D/2D array ARR to worksheet WSH in Excel spreadsheet
     file FILENAME in range RANGE.

     RSTATUS returns 1 if write succeeded, 0 otherwise.

     FILENAME must be a valid .xls Excel file name. If FILENAME does
     not contain any directory path, the file is saved in the current
     directory.

     ARR can be any array type save complex. Mixed numeric/text arrays
     can only be cell arrays.

     If only 3 arguments are given, the 3rd is assumed to be a
     spreadsheet range if it contains a ":" or is a completely empty
     string (corresponding to A1:IV65336). The 3rd argument is assumed
     to refer to a worksheet if it is a numeric value or a non-empty
     text string not containing ":"

     WSH can be a number or string (max. 31 chars).  In case of a not
     yet existing Excel file, the first worksheet will be used & named
     according to WSH - the extra worksheets that Excel creates by
     default are deleted.  In case of existing files, some checks are
     made for existing worksheet names or numbers, or whether WSH
     refers to an existing sheet with a type other than worksheet
     (e.g., chart).  When new worksheets are to be added to the Excel
     file, they are inserted to the right of all existing worksheets.
     The pointer to the "active" sheet (shown when Excel opens the
     file) remains untouched.

     RANGE is expected to be a regular spreadsheet range.  Data is
     added to the worksheet; existing data in the requested range will
     be overwritten.  Array ARR will be clipped at the right and/or
     bottom if its size is bigger than can be accommodated in RANGE.
     If ARR is smaller than the RANGE allows, it is placed in the top
     left of RANGE.

     If RANGE contains merged cells, only the elements of ARR
     corresponding to the top or left Excel cells of those merged cells
     will be written, other array cells corresponding to that cell will
     be ignored.

     The optional last argument REQINTF can be used to override the
     automatic selection by xlswrite of one interface out of the
     supported ones: COM/Excel, Java/Apache POI, or Java/JExcelAPI.
     For writing to OOXML files (.xlsx) a value of 'poi'
     (case-insensitive) must be specified for REQINTF.

     xlswrite is a mere wrapper for various scripts which find out what
     Excel interface to use (COM, Java/POI) plus code to mimic the other
     brand's syntax. For each call to xlswrite such an interface must be
     started and possibly an Excel file loaded. When writing to multiple
     ranges and/or worksheets in the same Excel file, a speed bonus can
     be obtained by invoking those scripts (xlsopen / octxls / .... /
     xlsclose) directly.

     Examples:

            status = xlswrite ('test4.xls', 'arr', 'Third_sheet', 'C3:AB40');
            (which adds the contents of array arr (any type) to range C3:AB40
            in worksheet 'Third_sheet' in file test4.xls and returns a logical
            True (= numerical 1) in status if al went well)

     See also: xlsread, oct2xls, xls2oct, xlsopen, xlsclose, xlsfinfo



# name: <cell-element>
# type: string
# elements: 1
# length: 80
Add data in 1D/2D array ARR to worksheet WSH in Excel spreadsheet file
FILENAME 

