Okay. This is a little awful, but you knew it was going to be.
First, you can access the actual format string for (for instance) 'L'
:
var formatL = moment.localeData().longDateFormat('L');
Next, you can perform some surgery on it with judicious regex replacement:
var formatYearlessL = formatL.replace(/Y/g,'').replace(/^\W|\W$|\W\W/,'');
(Which is to say: Remove YYYY, plus the orphaned separator left by its removal)
Then you can use your new format string in a moment format call:
someDate.format(formatYearlessL);
This necessarily makes some assumptions:
- The order of the month + day numeric format for a locale matches the order for the year + month + day format for that locale, with the year removed.
- The short form uses separators only between month and day (no leading / trailing separators).
- The separator for a short numeric date format is always non-alphanumeric.
- The format consists of numeric elements and separators, rather than a sentence-form format with articles (see RGPT's comment below about Spanish and Portugese, which will also apply to long formats in some other languages).
On a quick review of locale/*.js
, these assumptions hold true for every locale file I examined, but there may be some locales that violate them. (ETA: a comment below points out that a German short date format violates the second assumption)
As an additional important caveat, this is likely to be fragile. It is entirely possible that a future version of moment.js will change the location of the data currently in longDateFormat
...