I’m writing a PHP script to generate SQL dumps from my database for version control purposes. It already dumps the data structure by means of running the appropriate
SHOW CREATE .... query. Now I want to dump data itself but I’m unsure about the best method. My requirements are:
- I need a record per row
- Rows must be sorted by primary key
- SQL must be valid and exact no matter the data type (integers, strings, binary data…)
- Dumps should be identical when data has not changed
I can detect and run
mysqldump as external command but that adds an extra system requirement and I need to parse the output in order to remove headers and footers with dump information I don’t need (such as server version or dump date). I’d love to keep my script as simple as I can so it can be hold in an standalone file.
What are my alternatives?
I think the mysqldump output parsing is still the easiest. I think there are really few meta-data you have to exclude, and thus dropping those should be only a few lines of code.
You could also look at the following mysqdump options: –tab, –compact
Could always try the system() function. Hope you are in linux 🙂
“Select * from tablename order by primary key”
you can get the table names from the information_schema database or the “show tables” command. Also can get the primary key from “show indexes from tablename”
and then loop and create insert statements strings