I've read http://erack.org/blog/archives/8-LibreOffice-date-acceptance-patterns.html and I don't find this situation helpful. ISO 8601 is a standard and is not locale-independent. I don't care that much about the output/display format, but formats such as "2014-11-08T19:24:24" should always be recognised.
User story: as a user with a a list of ISO 8601 timestamp, I want them to be consistently recognised as timestamps, so that they are displayed as dates and can be worked on as dates (e.g. to calculate time intervals/differences).
Bug 72040 is related (or can be considered a blocker/subtask) but seems focused on a single ISO 8601 format and on display.
I'm using 126.96.36.199-1.fc20 with Italian locale. My specific input was https://frdata.wikimedia.org/campaign-vs-amount.csv
in your .csv file is format of date as in bug 72040. I think you have to specify which other format Calc should accept, because ISO 8601 allows reduced precision.
setting as enhancement
Your .csv file opens with ISO dates recognized as date whatever locale I'm using (my default is French) with the box "Detect special numbers" checked.
(In reply to Nemo_bis from comment #0)
> I've read
> and I don't find this situation helpful.
> User story: as a user with a a list of ISO 8601 timestamp, I want them to be
> consistently recognised as timestamps, so that they are displayed as dates
> and can be worked on as dates (e.g. to calculate time intervals/differences).
> I'm using 188.8.131.52-1.fc20 with Italian locale. My specific input was
So we have a test file. What are the repro steps to take and the RESULTS vs what you EXPECTED to see?
Status -> NEEDINFO
(set status back to UNCONFIRMED after you provide those repro steps)
As comment 2 states, the full ISO 8601 date/time format is recognized if the "Detect special numbers" checkbox in the import dialogue is checked on.
What maybe could be done is to detect the date/time also if that is unchecked, but that would require to apply the same exact cell format as people may want to export the file again and preserve the ISO format, for which at the moment the unchecked "special numbers" can be used.
Removing NEEDINFO, changing summary as this is not related to date acceptance patterns but only the CSV import.
Thanks for figuring it out and sorry for jumping to conclusions based on docs.
I'm still not sure what steps I'm supposed to follow, this is what I tried now:
0) LibreOffice 184.108.40.206-2.fc20; it-IT interface, locale and default document language
1) "libreoffice campaign-vs-amount.csv"
2) In the import window select UTF-8, check "detect special numbers", confirm
I. Observed: The table is there, as is in the CSV.
3) Apply a cell format of any "date" kind to the date columns
II. Expected: dates are displayed in the chosen format.
III. Observed: dates remain in their "native" CSV format.
4) Apply any date function to any of the date cells, e.g. DATA.DIFF(E4;I4)
III. Expected: the date function gives an output.
IV. Observed: Err:511
5) Repeat steps 3-4 after changing locale to fr-FR in Tools > Options > Language settings > Language > Locale (but not interface, nor default document language, nor this document's language).
V. Observed: the date patterns in the same options panel vary.
III and IV again.
Am I still doing something wrong?
Strange, import works with other locales (you could change in step 2 above), but not with it-IT. Also manual input of 2014-11-23T12:34:56 is not accepted as date in it-IT, but is in other locales.
That's a bug, I'll investigate.
Split-off bug 89500 for this.