Re: Insert statement changes timestamp value from MS Access

Tom,

Thanks for the quick reply and I'm sorry I wasn't explicit enough in my
description. At the top of the attachment there is this section which
is the piece of data from DD5 that is causing the problem:

****************
MS Access Details:

DATA TO IMPORT FROM LOCAL MSACCESS Table into linked ODBC pg table:
ID original_deployment_code date_time water_temp
depth_m salinity sp_cond do_per do_mg_l ph
1 MLM20010327 2001-04-01 02:00:00 17.860001
1.49352 30.1 46.299999 80.400002 6.42 7.64
2 MLM20010327 2001-04-01 02:30:00 17.700001
1.61544 30.200001 46.400002 78.699997 6.3 7.64
3 MLM20010327 2001-04-01 03:00:00 17.67
1.64592 30.200001 46.400002 72.800003 5.84 7.62
4 MLM20010327 2001-04-01 03:30:00 17.639999
1.524 30.1 46.299999 79.300003 6.36 7.61
********************

I needed to get this data in, so I tried to manually correct the invalid
date_time field after inserting and then put the uniqueness constraint
back on. From EMS I tried to change the times from 3:30 and 3:00 to
2:30 and 2:00. When I refresh the data it now shows 1:30 and 1:00 for
these records! I then tried to change 1:30 and 1:00 to 2:30 and 2:00
and it went back to 3:30 and 3:00.

So I backed up and tried to import directly from psql. The same results
happen as when I did it from MS Access.

I was driving home and realized 4/1/2001 was probably when daylight
savings time changed.....and it was. So I'm not sure if this is a bug
or not.

Regards,

Dave

Tom Lane wrote:

>"David Dabney" <David(dot)Dabney(at)noaa(dot)gov> writes:
>
>
>>I'm assuming this is a bug.
>>
>>
>
>Perhaps, but there is absolutely no chance of anyone reproducing the
>problem from the information you've offered. You say "it fails when
>I try to insert data from table DD5", but you give us no clue about
>what data is in DD5.
>
>There is some generic advice about how to create a useful bug report
>here:
>http://www.postgresql.org/docs/8.0/static/bug-reporting.html
>
>The short and sweet version of it is "give us a SQL script that
>delivers a wrong result" ...
>
> regards, tom lane
>
>---------------------------(end of broadcast)---------------------------
>TIP 2: Don't 'kill -9' the postmaster
>
>