I agree we should allow specifying timestamp format at column level, and respect that at data loading as well as query time. Is there technical reason which prevent us from doing that? Otherwise, respecting the specification should be the way to go.
col1(Date) col2(Date)
> > 2016-09-24 2016-09-25 00:00:00
Regards.
Jihong
-----Original Message-----
From: Venkata Gollamudi [mailto:
[hidden email]]
Sent: Monday, September 26, 2016 1:05 PM
To:
[hidden email]
Subject: Re: [Discussion] Support Date/Time format for Timestamp columns to be defined at column level
I completely agree with Vimal das.
On Sat, Sep 24, 2016, 10:05 PM Vimal Das Kammath <
[hidden email]>
wrote:
> Hi,
>
> The date format at column level should be to support loading data into
> carbon from csv files which have multiple columns with different date
> format. Currently data loading will fail if the multiple date columns are
> in different format.
>
> For the query result, my opinion is that the all the date columns in the
> result should be in a single format.
>
> Regards
> Vimal
>
>
> On Sep 24, 2016 2:11 PM, "向志强" <
[hidden email]> wrote:
>
> > Hi, all
> >
> > In recent days, I am trying to handle issues CARBONDATA-37. We are trying
> > to support that Date format can be set at column level.
> >
> > There is a doubt that we should feedback the the same format for Date
> > column or feedback a uniform format. Absolutely.
> >
> > For example.
> >
> > we create a table and define two cols which data type is Date. But the
> Date
> > format is different.
> >
> > col1(Date) col2(Date)
> > 2016-09-24 2016-09-25 00:00:00
> >
> > when querying, for two formats below, which should be returned?
> >
> > col1(Date) col2(Date)
> > 2016-09-24 2016-09-25 00:00:00
> >
> > or
> >
> > col1(Date) col2(Date)
> > 2016-09-24 00:00:00 2016-09-25 00:00:00
> >
> > if we set YYYY-MM-DD HH:MM:SS as default format.
> >
> >
> > Best wishes!
> >
>
Locked
