On Thu, Sep 20, 2018 at 6:09 AM amul sul <sulamul@gmail.com> wrote:
> On Thu, Sep 20, 2018, 3:22 AM Alexander Korotkov <a.korotkov@postgrespro.ru> wrote:
>> It's hard to understand whether it was expected, because it wasn't
>> properly documented. More important that it's the same behavior we
>> have before cf984672, and purpose of cf984672 was not to change this.
>>
>> But from the code comments, it's intentional. If you put digits or
>> text characters into format string, you can skip non-separator input
>> string characters. For instance you may do.
>>
>> # SELECT to_date('y18y12y2011', 'xDDxMMxYYYY');
>> to_date
>> ------------
>> 2011-12-18
>> (1 row)
>>
>> It's even more interesting that letters and digits are handled in
>> different manner.
>>
>> # SELECT to_date('01801202011', 'xDDxMMxYYYY');
>> ERROR: date/time field value out of range: "01801202011"
>> Time: 0,453 ms
>>
>> # SELECT to_date('01801202011', '9DD9MM9YYYY');
>> to_date
>> ------------
>> 2011-12-18
>> (1 row)
>>
>> So, letters in format string doesn't allow you to extract fields at
>> particular positions of digit sequence, but digits in format string
>> allows you to. That's rather undocumented, but from the code you can
>> get that it's intentional. Thus, I think it would be nice to improve
>> the documentation here. But I still propose to commit the patch I
>> propose to bring back unintentional behavior change in cf984672.
>
> Agreed, thanks for working on this.
Pushed, thanks.
------
Alexander Korotkov
Postgres Professional: http://www.postgrespro.com
The Russian Postgres Company