MySQL FROM_UNIXTIME and UNIX_TIMESTAMP functions in PostgreSQL

Janusz Słota — Thu, 11 Jun 2009 20:00:42 +0000

Personally I prefer PostgreSQL over MySQL DBMS. However there are some things I miss in PostgreSQL. Here is an advice on how to use MySQL FROM_UNIXTIME and UNIX_TIMESTAMP function in PostgreSQL DBMS.

FROM_UNIXTIME()

MySQL query:

mysql> SELECT FROM_UNIXTIME(123456789);

PostgreSQL equivalent:

postgres=> SELECT to_timestamp(123456789); -- with time zone
postgres=> SELECT to_timestamp(123456789)::timestamp; -- without time zone

In order to be able to run MySQL style query in PostgreSQL DBMS, you need to create this function:

CREATE OR REPLACE FUNCTION from_unixtime(integer) RETURNS timestamp AS '
	 SELECT to_timestamp($1)::timestamp AS result
' LANGUAGE 'SQL';

then you can run query like this (exactly the same as in MySQL):

postgres=> SELECT FROM_UNIXTIME(123456789);
    from_unixtime
---------------------
 1973-11-29 21:33:09
(1 row)

You can probably find (on the Internet) definitions with keyword abstime (i.e. $1::abstime::timestamp without time zone AS result).
This keyword is obsolete as described here, so IMHO it’s better to use the definition you see above.

UNIX_TIMESTAMP()

MySQL queries:

mysql> SELECT UNIX_TIMESTAMP();
mysql> SELECT UNIX_TIMESTAMP('1973-11-29 21:33:09');

PostgreSQL equivalents:

postgres=> SELECT EXTRACT(EPOCH FROM CURRENT_TIMESTAMP(0));
postgres=> SELECT EXTRACT(EPOCH FROM TIMESTAMP '1973-11-29 21:33:09');

Again – in order to be able to run MySQL style query in PostgreSQL DBMS, you need to create 3 functions:

-- no params
CREATE OR REPLACE FUNCTION unix_timestamp() RETURNS bigint AS '
	SELECT EXTRACT(EPOCH FROM CURRENT_TIMESTAMP(0))::bigint AS result;
' LANGUAGE 'SQL';
 
-- timestamp without time zone (i.e. 1973-11-29 21:33:09)
CREATE OR REPLACE FUNCTION unix_timestamp(timestamp) RETURNS bigint AS '
	SELECT EXTRACT(EPOCH FROM $1)::bigint AS result;
' LANGUAGE 'SQL';
 
-- timestamp with time zone (i.e. 1973-11-29 21:33:09+01)
CREATE OR REPLACE FUNCTION unix_timestamp(timestamp WITH time zone) RETURNS bigint AS '
	SELECT EXTRACT(EPOCH FROM $1)::bigint AS result;
' LANGUAGE 'SQL';

As noted here, prior to PostgreSQL 7.3, writing just timestamp was equivalent to timestamp with time zone. This was changed for SQL compliance.

Now you can run queries like this:

postgres=> SELECT UNIX_TIMESTAMP();
postgres=> SELECT UNIX_TIMESTAMP('1973-11-29 21:33:09');
postgres=> SELECT UNIX_TIMESTAMP('1973-11-29 21:33:09' AT TIME ZONE 'GMT');
postgres=> SELECT UNIX_TIMESTAMP('1973-11-29 21:33:09+01');

Tested on: PostgreSQL 8.1.11 (debian), PostgreSQL 8.2.7 (gentoo) and MySQL 5.0.70 (gentoo)

What’s the average length of an email address?

Janusz Słota — Tue, 05 May 2009 16:52:18 +0000

This question always comes up, when you have to design new table in database to store email addresses. Is VARCHAR(30) long enough? Or maybe you should use VARCHAR(255) or maybe TEXT as I’ve seen in some projects?

My live database, which I’ve used for this test contains 92298 valid email addresses, verified with the Pear Validate class with domain checking.

if(Validate::email($email,true))
{
    // add to database
    return true;
}
else
{
    // return error
    return false;
}

It means, if the email address is in this database – it is real and correct, otherwise the field is empty. So to make sure not to include empty fields, I’ve issued following query:

SELECT count(*), length(email) AS size FROM customer WHERE email LIKE '%@%' GROUP BY size ORDER BY size;

This query returns set of values, which I’ve used to create this chart (with OpenOffice.org Calc).

As you can see majority of email addresses are around 13 to 34 characters long. To store email addresses I’ve used VARCHAR (255) but as you can see VARCHAR(50) is enough.

The answer: VARCHAR(50) should do.

Janusz Slota

MySQL FROM_UNIXTIME and UNIX_TIMESTAMP functions in PostgreSQL

FROM_UNIXTIME()

UNIX_TIMESTAMP()

What’s the average length of an email address?