I'm currently querying a database with a timestamp column expressed in nanoseconds since 2000/01/01. For example, I have the value:
600393600000000000
Which can be converted to a date using the following code (written in Dart but just as an example):
DateTime epoch = new DateTime(2000, 1, 1);
var date = epoch.add(new Duration(microseconds:(600393600000000000 / 1000).floor()));
print(date);
This would print:
2019-01-10 00:00:00.000
My goal is to convert this into a postgresql timestampz through a query rather than using any application code. I know postgresql has the to_timestamp() function which can be used to create a timestamp from unix time but can't find a method for this.
Can anyone help?
Just as already mentioned, use the following construct:
to_timestamp(extract(epoch from '2000-01-01 00:00:00Z'::timestamptz) + (<your_nanos>::decimal / 1000000000))
...e.g.:
select to_timestamp(extract(epoch from '2000-01-01 00:00:00Z'::timestamptz) + (600393600123456789::decimal / 1000000000));
...yields:
to_timestamp
-------------------------------
2019-01-10 00:00:00.123457+00
(1 row)
Related
very new to the backend as well as all things postgresql, at the moment all ive been able to do is
SELECT * FROM nutrition WHERE timestamp >= (extract(epoch from now())::bigint * 1000) - 86400000 AND timestamp <= (extract(epoch from now())::bigint * 1000) + 86400000
in the frontend using js, im using Date.now() to store the timestamp in the DB.
timestamp is a column in my db thats logging the unix time in bigint format in which the food was logged. I want to get all the data from the current day from the hours beteen 12 AM midnight, and 11:59 PM. thanks.
for example, the last item i logged was last night at 10pm (1663995295337 UNIX time) so the data shouldnt include it.
show timezone returns;
America/Los_Angeles
Solution below --------------------------------------------------------------------
const today = new Date();
const beginningOfDay = today.setUTCHours(7, 0, 0, 0);
const endOfDay = today.setUTCHours(30, 59, 59, 99);
switch (method) {
case "GET":
try {
const text = `
select
*
from
nutrition
where
timestamp
between ${beginningOfDay} and ${endOfDay}`
this was the solution i was looking for, thanks for the help. sorry if i wasnt descriptive enough.
Assuming that by Unix time you mean epoch.
select extract(epoch from now());
extract
-------------------
1664038032.392004
select to_timestamp(1664038032.392004);
to_timestamp
--------------------------------
09/24/2022 09:47:12.392004 PDT
select
*
from
some_table
where
to_timestamp(1664038032.392004)
between
current_date + '00:00:00'::time AND current_date + '23:59:59'::time
UPDATE
Using timestamptz field in Postgres and an ISO datetime string from Javascript in order to properly deal with time zone.
create table tsz_test(id integer, tsz_fld timestamptz);
--In Web frontend today = new Date().toISOString(); "2022-09-24T20:57:05.830Z"
insert into
tsz_test
values (1, '2022-09-24T20:57:05.830Z'), (2, '2022-09-25T08:57:05.830Z');
select * from tsz_test ;
id | tsz_fld
----+----------------------------
1 | 09/24/2022 13:57:05.83 PDT
2 | 09/25/2022 01:57:05.83 PDT
--Borrowing from #a_horse_with_no_name answer
select * from tsz_test where tsz_fld::date = '09/24/2022'::date;
id | tsz_fld
----+----------------------------
1 | 09/24/2022 13:57:05.83 PDT
You can convert your "unix timestamp" to a date, then compare it with "today":
where to_timestamp("timestamp")::date = current_date
This assumes that your column named "timestamp" is not really a timestamp
If the column doesn't actually store seconds (which would be a unix epoch), but milliseconds you need to_timestamp("timestamp"/1000)::date instead (another source of problems that wouldn't exist if you had used a proper timestamptz or at least timestamp data type).
I want to query postgres database table which has the column with timestamp in long milliseconds. But I have the time in date format "yyyy-MM-dd HH:mm:ssZ" like this. How can I convert this date format to long milliseconds to run the query?
You can either convert your long value to a proper timestamp:
select *
from the_table
where to_timestamp(the_millisecond_column / 1000) = timestamp '2020-10-05 07:42'
Or extract the seconds from the timestamp value :
select *
from the_table
where the_millisecond_column = extract(epoch from timestamp '2020-10-05 07:42') * 1000
The better solution is however to convert that column to a proper timestamp column to avoid the constant conversion between (milliseconds) and proper timestamp values
I have a SELECT statement that I am trying to convert from Netezza SQL to Impala SQL. The output looks something like 140612, which is a date that is obtained by subtracting 7 from the current date and then pulling out the monday of that week.
I need to have this readable for Impala, then format it, then turn it into a string.
The query is :
TO_CHAR(next_day(DATE(a.date)-7, 'Monday'), 'YYMMDD') AS START_DATE
Assuming a.date is a timestamp, and T is the day of the week (1 = Sunday, 7 = Saturday; for your example above, Monday = 2, so T = 2) you should be able to use use
date_add(a.date, 7 - pmod(dayofweek(a.date) - T, 7));
in place of next_day in the above query. Check out the documentation on Impala's built-in date and time functions for more detail.
I am trying to retrieve time difference in minutes from a table(login_history as t1) using postgresql .
When i tried this code
((date_part('hour', timestamp '2014-04-25 09:44:21')- date_part('hour', timestamp '2014-04-25 08:32:21'))*60 +(date_part('minutes', timestamp '2014-04-25 09:44:21')- date_part('minutes', timestamp '2014-04-25 08:32:21'))) as TimeNew
It works fine.
But when i tried to retrieve information from a table t1 using this code
((date_part('hour', timestamp t1.login_date)- date_part('hour', timestamp t1.logout_date))*60 +
(date_part('minutes', timestamp t1.login_date)- date_part('minutes', timestamp t1.logout_date))
) as TimeNew
It throws this error
SQLSTATE[42601]: Syntax error: 7 ERROR: syntax error at or near "t1"
Thanks
I would use the interval that results from subtracting two timestamps for a much simpler expression:
select extract (epoch from (timestamp '2014-04-25 09:44:21' - timestamp '2014-04-25 08:32:21'))::integer/60
(gives 72)
or for your table:
select extract (epoch from (t1.logout_date - t1.login_date))::integer/60
If you need to cast:
select extract (epoch from (t1.logout_date::timestamp - t1.login_date::timestamp))::integer/60
or see the to_timestamp function for custom string parsing: http://www.postgresql.org/docs/9.4/static/functions-formatting.html
I needed to remove the timestamp from the query before t1 and the query works.
My first query is:
SELECT distinct wfc_request_job_id,wfc_request_job_info,
replace(iso_cc,';',' ') as "iso_cc",to_char(wfc_request_start_ts,'yyyy-MM-dd HH:mm:ss') as ts,
sent_message_count,
(link_object_count + poi_object_count + point_address_object_count) as request_object_count
FROM wfc_request_job
where
wfc_request_job_id=173526;
This returns ts as 2015-08-16 03:08:59
Second Query:
SELECT wfc_request_job_id,wfc_request_start_ts,wfc_request_end_ts,replace(iso_cc,';',' ') as "iso_ccs",sent_message_count,wfc_queue_name
FROM wfc_request_job
where
to_char(wfc_request_start_ts,'YYYY-MM-DD') >= to_char(to_date('08/16/2015','MM/DD/YYYY'),'YYYY-MM-DD')
and to_char(wfc_request_start_ts,'YYYY-MM-DD') <= to_char(to_date('08/16/2015','MM/DD/YYYY'),'YYYY-MM-DD')
order by wfc_request_job_id desc
This returns ts of the job id mentioned above as - "2015-08-16 15:58:59.809+02"
How can I make both the queries return ts in UTC+02 - i.e. same timezone
The data type of wfc_request_start_ts is - timestamp with timezone
I changed to queries to have the format HH24:MI:SS however that did not help. Please note that the webapp using these queries will be opened in both Germany and USA.
According to postgresql manual to_char there is TZ (and OF as of v9.4) template patterns for Date/Time formatting.
Therefore in query you need to add it so
postgres=# select to_char(now(),'yyyy-MM-dd HH24:mm:ss TZ');
to_char
------------------------
2015-08-19 12:08:56 CEST
(1 row)
Also, make sure you specify timezone when converting
so instead
to_date('08/16/2015','MM/DD/YYYY')
use
TIMESTAMP WITH TIME ZONE '2015-08-16 00:00:00+02';
in second query.