Question

I have a large table on a remote server and want to keep a copy on the local machine in a flat disk file. If I keep a record of a serial number locally read from the first row of the table then can I issue a SELECT command to return all of the rows in the table if the contents of the column at row id=0 <> my locally stored serial number?

Something like

SELECT * FROM 1_makes WHERE IF id@0<> local_serial

Hope that makes sense

delimiter $$

CREATE TABLE `1_makes` (
  `id` int(11) NOT NULL,
  `make` varchar(20) DEFAULT NULL,
  `allow_global_disc` tinyint(4) DEFAULT NULL,
  `home_text` text,
  `pack_home_text` text,
  `update_ref` int(11) DEFAULT NULL,
  PRIMARY KEY (`id`)
) ENGINE=InnoDB DEFAULT CHARSET=utf8$$

So what I want is that the entire tables is returned only if the contents of row with id=0 and update_ref is different to the locally stored value.

If I get results then I know the data has changed so I can save it to a local disk file and update the read update_ref

Looks like this is what I need

I have removed the column update_ref and put the update value in the make column @ id=0

SELECT * FROM 1_makes WHERE (SELECT IF(make <> ‘my update number’, 1, 0) FROM 1_makes WHERE id = 0)

This works, just wanted to ask what the numbers 1, 0 relate to in IF(make <> ‘my update number’, 1, 0)

Answer 1

Your case would probably be better served by issuing a separate query. The application would first query the value of the row where id=0, and then proceed based on that.

However, you could accomplish this with a subquery:

SELECT *
FROM 1_make
WHERE (SELECT IF(update_ref <> :my_number, 1, 0) FROM 1_make WHERE id = 0)

Or an INNER JOIN:

SELECT t1.*
FROM 1_make t1
INNER JOIN 1_make t2
  ON t2.id = 0 AND t2.update_ref <> :my_number

However, if you’re just testing to see if there are results, it’s best just to return the count rather than return the unused values:

SELECT COUNT(*)
FROM 1_make
WHERE (SELECT IF(update_ref <> :my_number, 1, 0) FROM 1_make WHERE id = 0)

Or perhaps even better, use LIMIT to stop after the first row:

SELECT id
FROM 1_make
WHERE (SELECT IF(update_ref <> :my_number, 1, 0) FROM 1_make WHERE id = 0)
LIMIT 1

Or even better, just use the return value of the IF:

SELECT IF(update_ref <> :my_number, 1, 0)
FROM 1_make
WHERE id = 0

Even better, why store a value for update_ref for every row? Why not move it to its own table:

SELECT IF(update_ref <> :my_number, 1, 0)
FROM my_settings

The IF statement return 0, which means false, or 1, which means true.

Finally, if the id is an incremental value, why not just search for rows where the id is larger than the stored last value instead:

SELECT id
FROM 1_make
WHERE id > :my_number
LIMIT 1

Also, this doesn’t account for updates. Only, inserts.

Consider adding a slave server that is replicated from the master server. Then, simply automate backing up the slave server. This is how most backups are done.