Hi Ertan,

I think following query works for you.

postgres=# select date_part('epoch', ('2017-11-01 14:47:45'::timestamp(0) - '2017-10-31 13:22:17'::timestamp(0))/3600/24);

date_part

-----------

1.059352

(1 row)

postgres=# select date_part('epoch', ( '2017-10-31 13:22:17'::timestamp(0)-'2017-11-01 14:47:45'::timestamp(0))/3600/24);

date_part

-----------

-1.059352

(1 row)

Best regards.

İyi çalışmalar.

Samed YILDIRIM

07.12.2017, 12:59, "Ertan Küçükoğlu" <ertan.kucukoglu@1nar.com.tr>:

Hello,

There is this Excel report which will be produced by an application. In
Excel, there is below equation
31/10/2017 15:05 - 31/10/2017 14:36:00 = 0:28:21
2017-10-31 13:22:17 - 2017-11-01 14:47:45 = 1/1/1900 01:25

That is very simple in PostgreSQL. Simply subtract two timestamp without
time zone fields and you have the result. However, Excel also represent that
result 0:28:21 as double notation 0.0196874999965075 and 1/1/1900 01:25 as
1,05935185185081.

I could not see any way to have same values using PostgreSQL query. I tried:
extract(epoch from time_field2) - extract(epoch from time_field1) and result
is 1701 and 5128 respectively.

Putting aside reasons as to why numbers are used instead of more human
understandable time format, I would like to learn if having same results as
Excel is possible.

Thanks & regards,
Ertan Küçükoğlu