Tabular is a Ruby library for reading, writing, and manipulating CSV, tab-delimited and Excel data.

I extracted it from production code. Still extracting it, actually. I need to read structured data and manipulate it via a common interface before persisting it with ActiveRecord.

Tabular is also handy for display table-like data. For example, I want to display a bike race’s results in HTML. I need to drop empty columns: sometimes there are points or times; sometimes not. I need find the most precise time to format all the times in the results correctly.

Much of the API is a copy of FasterCSV without the focus on CSV.

Import and display can be configured with Mappers and Renderers. It’s a OOP-heavy design that is fast and test-able.

Tabular can read Excel files if you add the spreadsheet gem to your project.

Install

sudo gem install tabular

(The gem is hosted on Gemcutter, not RubyForge)

Dependencies

For tab-delimited data: Ruby standard lib

For CSV: FasterCSV (fastercsv.rubyforge.org/) sudo gem install fastercsv

For Excel: Spreadsheet gem (spreadsheet.rubyforge.org/) sudo gem install spreadsheet

Examples

>> table = Table.read(“test/fixtures/sample.csv”) >> table.rows.size

> 4

Access Table Rows by index: >> table

And Row cells as a Hash: >> table[:last_name]

> “Willson”

Usage

Table.read assumes that .txt files are tab-delimited, .csv files are comma-delimited, and .xls files are Excel. It assumes that the first row is the header row, and normalizes the header to lower-case with underscores. E.g., “Last Name” becomes “last_name”.

Table.new accepts an Array of Arrays or an Array of Hashes.

Table.new also accepts an options hash.

:columns option to map columns to a different key or type: :city_state => :location – Maps :city_state column to :location. A column with a “City State” header would be accessed as row :flyer_approved => { :column_type => :boolean } – Coerce :flyer_approved column cells to booleans.

:as => [:csv, :xls, :txt] to override file format

Tests

There’s basic test coverage. More comprehensive test coverage needs to be extracted from original projects. Run ‘rake test’.

Changes

0.2.7 Add Table#to_space_delimited for space-padded fixed layout 0.2.6 Add :except option for delete_blank_columns! 0.2.5 Use modern gemspec with no runtime dependencies. Make spreadsheet gem optional. 0.2.3 Add :except option for delete_homogenous_columns! 0.2.1 Documentation! 0.2.0 Add several new features that break previous API

* New public accessors for Table, Columns, Row, and Column
* Mapper to translate source data to Rows
* Renderer to control display of Row cells and Column headers
* Table#delete_blank_columns! to delete columns that are blank. Zero is considered blank.
* Table#delete_homogenous_columns! to delete columns that are all the same value. E.g.,
  A | B | C
  =========
  1 | 2 | 3
  1 | 6 |
  1 | * | 5

  Column A would be deleted
* Table#strip! to remove whitespace around cell values. By default, Tabular::Table preserves cell whitespace.
* Column#max
* Column#precision
* Ruby 1.8 support is deprecated

0.0.5 Parse ‘invalid’ m/d/yy dates