This plugin enables you to avoid rounding errors in your SQL select statements when selecting with conditions on a float column.

For example, if you have a row with column_a equal to 3.8, find_by_column_a(3.8) will return 0 results. This plugin will fix that.

acts_as_accurate_float works by overriding the default find statement to round off the stored float values before comparing them to your input values so that slight rounding errors don’t mess up your results.

find_by_column_a(3.8) will now produce the following SQL:

  SELECT * FROM `table_name` WHERE (round(column_a*100)/100 = 3.8) LIMIT 1;

It should work with all various find statements that include a condition on your acts_as_accurate_float field.

INSTALL

  script/plugin install git://github.com/jfredson/acts_as_accurate_float.git

Usage

  # specify a float column and level of accuracy you need
  # accuracy is specified in number of decimal places

  acts_as_accurate_float :column_a, 2

Compatibility

I have only verified these float issues with MySQL, and also have only tested this plugin for use with MySQL. If someone wants to take a crack at other databases, please let me know.

Copyright © 2009 John Fredrickson, released under the MIT license