Disregard percent signs in data comparisons

[mcoles]

Some CSV numerical data comes in with a percent sign at the end, and is treated as text. This causes issues when the values are the same but are formatted slightly differently due to floating point math (for which we have built-in rounding). If the column values are numeric except a "%" sign at the end, we should consider them numbers.