msqldump and version control systems

**chicane** · 17 February 2010, 17:47

Not sure if I've got the wrong end of the stick here, but have you tried the...

Code:

--skip-extended-insert

...option to insert one row per line?

The table structure should sit on multiple lines by default - if not you might want to look at the format of carriage returns/line feeds within the dump file to see if they match the format used by your client.

**NickFitz** · 17 February 2010, 18:04

--extended-insert=false will put each record into a separate INSERT.

EDIT: or wot chicane said

**Sysman** · 18 February 2010, 12:18

Thanks ya'll.

**oversteer** · 18 February 2010, 12:22

You can dump table structure and table data to separate files (-d and -t). Might be worth dumping each table to separate files with -T. Don't forget stored procedures, triggers etc..

**suityou01** · 18 February 2010, 22:49

Originally posted by Sysman View Post

The aim:

During development I want to dump MySQL tables in a format suitable for input to a version control system (Thanks Nick).

Plain mysqldump outputs all the data on one line, and there will be changes to the table structure along the way as well, so I am thinking that what I want is:

a dump of the table structure
a dump of the data with one line per record

To be honest, before we even talk about version control, being able to do a diff at the (bash) command line would be rather useful.

I came across this blog entry, but it relates to MySQL 4.n and I'm now on 5.1.42.

I have a FreeBSD web server and OS X as a client (also with MySQL) to go at.

As a sanity check, am I going in the right direction here?

I have a bash script that extracts all the tables and tarballs them up. PM me I'll send it to you. Might be some stuff of use to pick the bones out of.

msqldump and version control systems