SYNOPSIS
use Spreadsheet::Read;
my $book = ReadData ("test.csv", sep => ";");
my $book = ReadData ("test.sxc");
my $book = ReadData ("test.ods");
my $book = ReadData ("test.xls");
my $book = ReadData ("test.xlsx");
my $book = ReadData ($fh, parser => "xls");
my $sheet = $book->[1]; # first datasheet
my $cell = $book->[1]{A3}; # content of field A3 of sheet 1
my $cell = $book->[1]{cell}[1][3]; # same, unformatted
DESCRIPTION
Spreadsheet::Read tries to transparently read *any* spreadsheet and return its content in a universal manner independent of the parsing module that does the actual spreadsheet scanning.For OpenOffice and/or LibreOffice this module uses Spreadsheet::ReadSXC <http://metacpan.org/release/Spreadsheet-ReadSXC>
For Microsoft Excel this module uses Spreadsheet::ParseExcel <http://metacpan.org/release/Spreadsheet-ParseExcel>, Spreadsheet::ParseXLSX <http://metacpan.org/release/Spreadsheet-ParseXLSX>, or Spreadsheet::XLSX <http://metacpan.org/release/Spreadsheet-XLSX> (discouraged).
For CSV this module uses Text::CSV_XS <http://metacpan.org/release/Text-CSV_XS> or Text::CSV_PP <http://metacpan.org/release/Text-CSV_PP>.
For SquirrelCalc there is a very simplistic built-in parser
Data structure
The data is returned as an array reference:
$book = [
# Entry 0 is the overall control hash
{ sheets => 2,
sheet => {
"Sheet 1" => 1,
"Sheet 2" => 2,
},
type => "xls",
parser => "Spreadsheet::ParseExcel",
version => 0.59,
error => undef,
},
# Entry 1 is the first sheet
{ label => "Sheet 1",
maxrow => 2,
maxcol => 4,
cell => [ undef,
[ undef, 1 ],
[ undef, undef, undef, undef, undef, "Nugget" ],
],
attr => [],
merged => [],
A1 => 1,
B5 => "Nugget",
},
# Entry 2 is the second sheet
{ label => "Sheet 2",
:
:
To keep as close contact to spreadsheet users, row and column 1 have index 1 too in the "cell" element of the sheet hash, so cell ``A1'' is the same as "cell" [1, 1] (column first). To switch between the two, there are two helper functions available: "cell2cr ()" and "cr2cell ()".
The "cell" hash entry contains unformatted data, while the hash entries with the traditional labels contain the formatted values (if applicable).
The control hash (the first entry in the returned array ref), contains some spreadsheet meta-data. The entry "sheet" is there to be able to find the sheets when accessing them by name:
my %sheet2 = %{$book->[$book->[0]{sheet}{"Sheet 2"}]};
Functions
ReadData
my $book = ReadData ($source [, option => value [, ... ]]);
my $book = ReadData ("file.csv", sep => ',', quote => '"');
my $book = ReadData ("file.xls", dtfmt => "yyyy-mm-dd");
my $book = ReadData ("file.ods");
my $book = ReadData ("file.sxc");
my $book = ReadData ("content.xml");
my $book = ReadData ($content);
my $book = ReadData ($fh, parser => "xls");
Tries to convert the given file, string, or stream to the data structure described above.
Processing Excel data from a stream or content is supported through a File::Temp <https://metacpan.org/release/File-Temp> temporary file or IO::Scalar <https://metacpan.org/release/IO-Scalar> when available.
Spreadsheet::ReadSXC <https://metacpan.org/release/Spreadsheet-ReadSXC> does preserve sheet order as of version 0.20.
Currently supported options are:
- parser
-
Force the data to be parsed by a specific format. Possible values are
"csv", "prl" (or "perl"), "sc" (or "squirelcalc"), "sxc" (or "oo",
"ods", "openoffice", "libreoffice") "xls" (or "excel"), and "xlsx"
(or "excel2007").
When parsing streams, instead of files, it is highly recommended to pass this option.
Spreadsheet::Read supports several underlying parsers per spreadsheet type. It will try those from most favored to least favored. When you have a good reason to prefer a different parser, you can set that in environment variables. The other options then will not be tested for:
env SPREADSHEET_READ_CSV=Text::CSV_PP ...
- cells
- Control the generation of named cells (""A1"" etc). Default is true.
- rc
- Control the generation of the {cell}[c][r] entries. Default is true.
- attr
- Control the generation of the {attr}[c][r] entries. Default is false. See ``Cell Attributes'' below.
- clip
-
If set, "ReadData" will remove all trailing rows and columns
per sheet that have no visual data. If a sheet has no data at all, the
sheet will be skipped entirely when this attribute is true.
This option is only valid if "cells" is true. The default value is true if "cells" is true, and false otherwise.
- strip
-
If set, "ReadData" will remove trailing- and/or
leading-whitespace from every field.
strip leading strailing ----- ------- --------- 0 n/a n/a 1 strip n/a 2 n/a strip 3 strip strip - sep
- Set separator for CSV. Default is comma ",".
- quote
- Set quote character for CSV. Default is """.
- dtfmt
-
Set the format for MS-Excel date fields that are set to use the default
date format. The default format in Excel is ""m-d-yy"``, which is both
not year 2000 safe, nor very useful. The default is now ''"yyyy-mm-dd"",
which is more ISO-like.
Note that date formatting in MS-Excel is not reliable at all, as it will store/replace/change the date field separator in already stored formats if you change your locale settings. So the above mentioned default can be either ""m-d-yy"`` OR ''"m/d/yy"" depending on what that specific character happened to be at the time the user saved the file.
- debug
-
Enable some diagnostic messages to STDERR.
The value determines how much diagnostics are dumped (using Data::Peek <https://metacpan.org/release/Data-Peek>). A value of 9 and higher will dump the entire structure from the back-end parser.
All other attributes/options will be passed to the underlying parser if that parser supports attributes.
cr2cell
my $cell = cr2cell (col, row)
"cr2cell ()" converts a "(column, row)" pair (1 based) to the traditional cell notation:
my $cell = cr2cell ( 4, 14); # $cell now "D14" my $cell = cr2cell (28, 4); # $cell now "AB4"
cell2cr
my ($col, $row) = cell2cr ($cell)
"cell2cr ()" converts traditional cell notation to a "(column, row)" pair (1 based):
my ($col, $row) = cell2cr ("D14"); # returns ( 4, 14)
my ($col, $row) = cell2cr ("AB4"); # returns (28, 4)
row
my @row = row ($sheet, $row) my @row = Spreadsheet::Read::row ($book->[1], 3)
Get full row of formatted values (like "$sheet->{A3} .. $sheet->{G3}")
Note that the indexes in the returned list are 0-based.
"row ()" is not imported by default, so either specify it in the use argument list, or call it fully qualified.
cellrow
my @row = cellrow ($sheet, $row) my @row = Spreadsheet::Read::cellrow ($book->[1], 3)
Get full row of unformatted values (like "$sheet->{cell}[1][3] .. $sheet->{cell}[7][3]")
Note that the indexes in the returned list are 0-based.
"cellrow ()" is not imported by default, so either specify it in the use argument list, or call it fully qualified.
rows
my @rows = rows ($sheet) my @rows = Spreadsheet::Read::rows ($book->[1])
Convert "{cell}"'s "[column][row]" to a "[row][column]" list.
Note that the indexes in the returned list are 0-based, where the index in the "{cell}" entry is 1-based.
"rows ()" is not imported by default, so either specify it in the use argument list, or call it fully qualified.
parses
parses ($format)
Spreadsheet::Read::parses ("CSV")
"parses ()" returns Spreadsheet::Read's capability to parse the required format. "ReadData" will pick its preferred parser for that format unless overruled. See "parser".
"parses ()" is not imported by default, so either specify it in the use argument list, or call it fully qualified.
Version
my $v = Version () my $v = Spreadsheet::Read::Version () my $v = Spreadsheet::Read->VERSION;
Returns the current version of Spreadsheet::Read.
"Version ()" is not imported by default, so either specify it in the use argument list, or call it fully qualified.
This function returns exactly the same as "Spreadsheet::Read->VERSION" returns and is only kept for backward compatibility reasons.
Using CSV
In case of CSV parsing, "ReadData" will use the first line of the file to auto-detect the separation character if the first argument is a file and both "sep" and "quote" are not passed as attributes. Text::CSV_XS <https://metacpan.org/release/Text-CSV_XS> (or Text::CSV_PP <https://metacpan.org/release/Text-CSV_PP>) is able to automatically detect and use "\r" line endings.CSV can parse streams too, but be sure to pass "sep" and/or "quote" if these do not match the default "," and """.
When an error is found in the CSV, it is automatically reported (to STDERR). The structure will store the error in "$ss->[0]{error}" as anonymous list returned by "$csv->error_diag" <https://metacpan.org/pod/Text::CSV_XS#error_diag>. See Text::CSV_XS <https://metacpan.org/pod/Text-CSV_XS> for documentation.
my $ss = ReadData ("bad.csv");
$ss->[0]{error} and say $ss->[0]{error}[1];
Cell Attributes
If the constructor was called with "attr" having a true value, effort is made to analyze and store field attributes like this:
{ label => "Sheet 1",
maxrow => 5,
maxcol => 2,
cell => [ undef,
[ undef, 1 ],
[ undef, undef, undef, undef, undef, "Nugget" ],
],
attr => [ undef,
[ undef, {
type => "numeric",
fgcolor => "#ff0000",
bgcolor => undef,
font => "Arial",
size => undef,
format => "## ##0.00",
halign => "right",
valign => "top",
uline => 0,
bold => 0,
italic => 0,
wrap => 0,
merged => 0,
hidden => 0,
locked => 0,
enc => "utf-8",
}, ]
[ undef, undef, undef, undef, undef, {
type => "text",
fgcolor => "#e2e2e2",
bgcolor => undef,
font => "Letter Gothic",
size => 15,
format => undef,
halign => "left",
valign => "top",
uline => 0,
bold => 0,
italic => 0,
wrap => 0,
merged => 0,
hidden => 0,
locked => 0,
enc => "iso8859-1",
}, ]
merged => [],
A1 => 1,
B5 => "Nugget",
},
This has now been partially implemented, mainly for Excel, as the other parsers do not (yet) support all of that. YMMV.
If a cell itself is not hidden, but the parser holds the information that either the row or the column (or both) the field is in is hidden, the flag is inherited into the cell attributes.
Merged cells
Note that only Spreadsheet::ReadSXC <http://metacpan.org/release/Spreadsheet-ReadSXC> documents the use of merged cells, and not in a way useful for the spreadsheet consumer.
CSV does not support merged cells (though future implementations of CSV for the web might).
The documentation of merged areas in Spreadsheet::ParseExcel <http://metacpan.org/release/Spreadsheet-ParseExcel> and Spreadsheet::ParseXLSX <http://metacpan.org/release/Spreadsheet-ParseXLSX> can be found in Spreadsheet::ParseExcel::Worksheet <http://metacpan.org/release/Spreadsheet-ParseExcel-Worksheet> and Spreadsheet::ParseExcel::Cell <http://metacpan.org/release/Spreadsheet-ParseExcel-Cell>.
None of basic Spreadsheet::XLSX <http://metacpan.org/release/Spreadsheet-XLSX>, Spreadsheet::ParseExcel <http://metacpan.org/release/Spreadsheet-ParseExcel>, and Spreadsheet::ParseXLSX <http://metacpan.org/release/Spreadsheet-ParseXLSX> manual pages mention merged cells at all.
This module just tries to return the information in a generic way.
Given this spreadsheet as an example
merged.xlsx:
A B C
+-----+-----------+
1| | foo |
+-----+ +
2| bar | |
| +-----+-----+
3| | urg | orc |
+-----+-----+-----+
the information extracted from that undocumented information is returned in the "merged" entry of the sheet's hash as a list of top-left, bottom-right coordinate pars (col, row, col, row). For given example, that would be:
$ss->{merged} = [
[ 1, 2, 1, 3 ], # A2-A3
[ 2, 1, 3, 2 ], # B1-C2
];
When the attributes are also enabled, there is some merge information copied directly from the cell information, but again, that stems from code analysis and not from documentation:
my $ss = ReadData ("merged.xlsx", attr => 1)->[1];
foreach my $row (1 .. $ss->{maxrow}) {
foreach my $col (1 .. $ss->{maxcol}) {
my $cell = cr2cell ($col, $row);
printf "%s %-3s %d ", $cell, $ss->{$cell},
$ss->{attr}[$col][$row]{merged};
}
print "\n";
}
A1 0 B1 foo 1 C1 1
A2 bar 1 B2 1 C2 1
A3 1 B3 urg 0 C3 orc 0
In this example, there is no way to see if "B2" is merged to "A2" or to "B1" without analyzing all surrounding cells. This could as well mean "A2:A3", "B1:C1", "B2:C2", as "A2:A3", "B1:B2", "C1:C2", as "A2:A3", "B1:C2". Use the "merged" entry described above to find out what fields are merged to what other fields.
TOOLS
This modules comes with a few tools that perform tasks from the FAQ, like ``How do I select only column D through F from sheet 2 into a CSV file?''
If the module was installed without the tools, you can find them here:
https://github.com/Tux/Spreadsheet-Read/tree/master/examples
xlscat
Show (parts of) a spreadsheet in plain text, CSV, or HTML
usage: xlscat [-s <sep>] [-L] [-n] [-A] [-u] [Selection] file.xls
[-c | -m] [-u] [Selection] file.xls
-i [-S sheets] file.xls
Generic options:
-v[#] Set verbose level (xlscat/xlsgrep)
-d[#] Set debug level (Spreadsheet::Read)
-u Use unformatted values
--noclip Do not strip empty sheets and
trailing empty rows and columns
-e <enc> Set encoding for input and output
-b <enc> Set encoding for input
-a <enc> Set encoding for output
Input CSV:
--in-sep=c Set input sep_char for CSV
Input XLS:
--dtfmt=fmt Specify the default date format to replace 'm-d-yy'
the default replacement is 'yyyy-mm-dd'
Output Text (default):
-s <sep> Use separator <sep>. Default '|', \n allowed
-L Line up the columns
-n [skip] Number lines (prefix with column number)
optionally skip <skip> (header) lines
-A Show field attributes in ANSI escapes
-h[#] Show # header lines
Output Index only:
-i Show sheet names and size only
Output CSV:
-c Output CSV, separator = ','
-m Output CSV, separator = ';'
Output HTML:
-H Output HTML
Selection:
-S <sheets> Only print sheets <sheets>. 'all' is a valid set
Default only prints the first sheet
-R <rows> Only print rows <rows>. Default is 'all'
-C <cols> Only print columns <cols>. Default is 'all'
-F <flds> Only fields <flds> e.g. -FA3,B16
Ordering (column numbers in result set *after* selection):
--sort=spec Sort output (e.g. --sort=3,2r,5n,1rn+2)
+# - first # lines do not sort (header)
# - order on column # lexical ascending
#n - order on column # numeric ascending
#r - order on column # lexical descending
#rn - order on column # numeric descending
xlsgrep
Show (parts of) a spreadsheet that match a pattern in plain text, CSV, or HTML
usage: xlsgrep [-s <sep>] [-L] [-n] [-A] [-u] [Selection] pattern file.xls
[-c | -m] [-u] [Selection] pattern file.xls
-i [-S sheets] pattern file.xls
Generic options:
-v[#] Set verbose level (xlscat/xlsgrep)
-d[#] Set debug level (Spreadsheet::Read)
-u Use unformatted values
--noclip Do not strip empty sheets and
trailing empty rows and columns
-e <enc> Set encoding for input and output
-b <enc> Set encoding for input
-a <enc> Set encoding for output
Input CSV:
--in-sep=c Set input sep_char for CSV
Input XLS:
--dtfmt=fmt Specify the default date format to replace 'm-d-yy'
the default replacement is 'yyyy-mm-dd'
Output Text (default):
-s <sep> Use separator <sep>. Default '|', \n allowed
-L Line up the columns
-n [skip] Number lines (prefix with column number)
optionally skip <skip> (header) lines
-A Show field attributes in ANSI escapes
-h[#] Show # header lines
Grep options:
-i Ignore case
-w Match whole words only
Output CSV:
-c Output CSV, separator = ','
-m Output CSV, separator = ';'
Output HTML:
-H Output HTML
Selection:
-S <sheets> Only print sheets <sheets>. 'all' is a valid set
Default only prints the first sheet
-R <rows> Only print rows <rows>. Default is 'all'
-C <cols> Only print columns <cols>. Default is 'all'
-F <flds> Only fields <flds> e.g. -FA3,B16
Ordering (column numbers in result set *after* selection):
--sort=spec Sort output (e.g. --sort=3,2r,5n,1rn+2)
+# - first # lines do not sort (header)
# - order on column # lexical ascending
#n - order on column # numeric ascending
#r - order on column # lexical descending
#rn - order on column # numeric descending
xls2csv
Convert a spreadsheet to CSV. This is just a small wrapper over "xlscat".
usage: xls2csv [ -o file.csv ] file.xls
ss2tk
Show a spreadsheet in a perl/Tk spreadsheet widget
usage: ss2tk [-w <width>] [X11 options] file.xls [<pattern>]
-w <width> use <width> as default column width (4)
ssdiff
Show the differences between two spreadsheets.
usage: examples/ssdiff [--verbose[=1]] file.xls file.xlsx
TODO
- Options
-
-
- Module Options
- New Spreadsheet::Read options are bound to happen. I'm thinking of an option that disables the reading of the data entirely to speed up an index request (how many sheets/fields/columns). See "xlscat -i".
- Parser options
- Try to transparently support as many options as the encapsulated modules support regarding (un)formatted values, (date) formats, hidden columns rows or fields etc. These could be implemented like "attr" above but names "meta", or just be new values in the "attr" hashes.
-
- Other spreadsheet formats
- I consider adding any spreadsheet interface that offers a usable API.
- Alternative parsers for existing formats
- As long as the alternative has a good reason for its existence, and the API of that parser reasonable fits in my approach, I will consider to implement the glue layer, or apply patches to do so as long as these match what CONTRIBUTING.md describes.
- Add an OO interface
- Consider making the ref an object, though I currently don't see the big advantage (yet). Maybe I'll make it so that it is a hybrid functional / OO interface.
AUTHOR
H.Merijn Brand, <[email protected]>COPYRIGHT AND LICENSE
Copyright (C) 2005-2016 H.Merijn BrandThis library is free software; you can redistribute it and/or modify it under the same terms as Perl itself.