=head1 ABOUT This archive contains the distribution Text-NumericData It consists of some modules for parsing and working with files that contain columns of numerical data in text files (ASCII, dat, txt, csv, tsv ... I name them txd, for "TeXt numerical Data"). Those files occur as a flexible human-readable format for exchange of small to intermediate data sets. As such, they appear in rich variety and need some flexibility in programs that work with them. The main benefit of the package is the supplied set of tools to directly work with these files, offering various operations like selection of subsets, sorting, basic statistics and flexible computations (also involving auxillary files influencing the main data set). Those operations are ideally combined via pipelines constructed in the shell (a cross-platform GUI for building filtering pipelines exists for an earlier version and might re-emerge if demand is there). The tools, starting with the prefix "txd", generally work as filters and are designed to work on multiple files sent via L, to avoid the penalty of recompiling the scripts all the time --- the raw parsing and computing performance is secondary to that, for usual data sets. There are some specific tools to convert from more esoteric formats. While there is some general adaptivity in the parsing, this can be necessary to get the data into an understandable shape. To get an idea of the canonical format the package uses, you can fire up txdconstruct --numformat=%.3f '[1]=C2; [2]=sin(2*pi()*[1])' \ | txdfilter 'The title' x y which produces this content: #The title #"x" "y" 0.000 0.000 0.111 0.643 0.222 0.985 0.333 0.866 0.444 0.342 0.556 -0.342 0.667 -0.866 0.778 -0.985 0.889 -0.643 1.000 -0.000 Care has been taken to enable transparent handling of files with differing line endings (DOS and UNIX, theoretically old-style Mac, too, but only on old-style Macs) independent of the current platform. The rule is that output shares the line end of the input, unless you override it on the command line. Handling of command-line parameters and configuration files as well as the actual generation of POD for the tools is achieved via L. This package starts with version 2, indicating that it is the successor to the TextData package by the same author, which was not available via CPAN. About eight years of experience went into straightening things out. The main changes are in the module API, while the command-line tools are mostly compatible to their predecessors. You should test your use-case and adapt the odd changed parameter (or, more sober script name in the case of txdderiv, not being txderiv anymore) when upgrading from TextData. =head1 INSTALLATION To install this module, run the following commands: perl Build.PL ./Build ./Build test ./Build install =head1 SUPPORT AND DOCUMENTATION After installing, you can find documentation for the main module with the perldoc command. perldoc Text::NumericData The installed scripts with the prefix txd also feature POD, which is installed as man-pages and give the same information interactively, for example txdcalc with txdcalc --help Have fun. =head1 LICENSE AND COPYRIGHT Copyright (C) 2005-2013 Thomas Orgis This is free software; you can redistribute it and/or modify it under the same terms as the Perl 5 programming language system itself.