Number formatting with “d” code – needs explanation & example fix

[globalizejs]

Number formatting with “d” code – needs explanation & example fix

[rdworth]

It seems like it doesn't work as expected (as documented) when a float is provided versus when an int is provided. Assuming it originated from .Net (which is where the culture info in Globalize originates from), it may not be intended for use with floats at all. See http://msdn.microsoft.com/en-us/library/dwhawy9k.aspx#DFormatString "This format is supported only for integral types."

It also seems to not handle negative numbers:

Globalize.format( -1234, "d7" ) // expected: "-0001234", actual "-1234"

thought the .Net version does.

I would not except "d" to provide grouping as that is provided by "n". So if you want "1,234" or "1 234" then call format with "n" but if you want "1234", call format with "d". Or for "01234", "d5". I think you're right in that it doesn't seem that will ever be affected by a culture, as it excludes grouping and thousands and decimal (point) separators.

[rdworth]

Added (failing) unit test in 47fc9e3 for negative integer with expected padding 0s after the negative sign.

[yucca42]

I think this is a documentation issue. Using plain “d” is rather pointless, as with it, format() does just the job of Number.toString(). With a numeric specifier, it’s potentially useful when you need fixed-width integers left padded with zeroes (though I wonder how common the need is, except for dates and times, for which it can be done another way).

I’d suggest changing the documentation to something like the following:

• d for number in non-localized format (as per Number.toString()) padded with zeros on the left

In the prose after the bulleted list, change
“The number determines how many decimal places to display for all the format types except decimal, for which it means the minimum number of digits to display, zero padding it if necessary.”
to
“The number determines how many decimal places to display for all the format types except ‘d’, for which it means the minimum number of characters to display, zero padding it on the left if necessary.”

Oh, wait... for a negative number, we always get the result as per Number.toString(), without zero padding, apparently because the code first takes the absolute value, then creates the formatted string (possibly with leading zeros), and finally does number = -number if the value was negative. And this destroys the zero padding.